{
  "best_metric": 0.12475299090147018,
  "best_model_checkpoint": "/home/datawork-iot-nos/Seatizen/models/multilabel/fine_scale/DinoVdeau-base-2024_09_03-batch-size32_epochs150_freeze/checkpoint-27573",
  "epoch": 111.0,
  "eval_steps": 500,
  "global_step": 30303,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.2079002079002079,
      "eval_f1_macro": 0.5105390450682302,
      "eval_f1_micro": 0.73108765167112,
      "eval_loss": 0.17516958713531494,
      "eval_roc_auc": 0.8186965528786462,
      "eval_runtime": 453.901,
      "eval_samples_per_second": 6.358,
      "eval_steps_per_second": 0.2,
      "learning_rate": 0.001,
      "step": 273
    },
    {
      "epoch": 1.8315018315018317,
      "grad_norm": 0.2602289021015167,
      "learning_rate": 0.001,
      "loss": 0.2857,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.23492723492723494,
      "eval_f1_macro": 0.5498069096094584,
      "eval_f1_micro": 0.7582569600553347,
      "eval_loss": 0.1577771008014679,
      "eval_roc_auc": 0.8363419499741919,
      "eval_runtime": 442.4314,
      "eval_samples_per_second": 6.523,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 546
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.23146223146223147,
      "eval_f1_macro": 0.6037272443934714,
      "eval_f1_micro": 0.7721545657578696,
      "eval_loss": 0.15162432193756104,
      "eval_roc_auc": 0.8505384953411333,
      "eval_runtime": 429.5087,
      "eval_samples_per_second": 6.719,
      "eval_steps_per_second": 0.212,
      "learning_rate": 0.001,
      "step": 819
    },
    {
      "epoch": 3.663003663003663,
      "grad_norm": 0.24485518038272858,
      "learning_rate": 0.001,
      "loss": 0.1764,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.24220374220374222,
      "eval_f1_macro": 0.613953187695023,
      "eval_f1_micro": 0.7649537378914902,
      "eval_loss": 0.15218119323253632,
      "eval_roc_auc": 0.8386795656946878,
      "eval_runtime": 440.157,
      "eval_samples_per_second": 6.557,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 1092
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.24220374220374222,
      "eval_f1_macro": 0.6161642626912543,
      "eval_f1_micro": 0.7719928186714542,
      "eval_loss": 0.14836864173412323,
      "eval_roc_auc": 0.840338176952158,
      "eval_runtime": 432.0583,
      "eval_samples_per_second": 6.68,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.001,
      "step": 1365
    },
    {
      "epoch": 5.4945054945054945,
      "grad_norm": 0.17316196858882904,
      "learning_rate": 0.001,
      "loss": 0.1677,
      "step": 1500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2560637560637561,
      "eval_f1_macro": 0.6051867487843677,
      "eval_f1_micro": 0.775030471878809,
      "eval_loss": 0.14818257093429565,
      "eval_roc_auc": 0.8434755477910759,
      "eval_runtime": 432.9746,
      "eval_samples_per_second": 6.666,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.001,
      "step": 1638
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.24185724185724186,
      "eval_f1_macro": 0.617739220969942,
      "eval_f1_micro": 0.7729166666666668,
      "eval_loss": 0.1486394852399826,
      "eval_roc_auc": 0.8431254755177443,
      "eval_runtime": 432.8426,
      "eval_samples_per_second": 6.668,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.001,
      "step": 1911
    },
    {
      "epoch": 7.326007326007326,
      "grad_norm": 0.31019529700279236,
      "learning_rate": 0.001,
      "loss": 0.1652,
      "step": 2000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.2512127512127512,
      "eval_f1_macro": 0.6171646674895677,
      "eval_f1_micro": 0.7767065175472426,
      "eval_loss": 0.14861202239990234,
      "eval_roc_auc": 0.8485322128731306,
      "eval_runtime": 436.0215,
      "eval_samples_per_second": 6.619,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 2184
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.2512127512127512,
      "eval_f1_macro": 0.6366264906922544,
      "eval_f1_micro": 0.7805490458654168,
      "eval_loss": 0.14834754168987274,
      "eval_roc_auc": 0.857034765243127,
      "eval_runtime": 443.0125,
      "eval_samples_per_second": 6.514,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.001,
      "step": 2457
    },
    {
      "epoch": 9.157509157509157,
      "grad_norm": 0.1974957138299942,
      "learning_rate": 0.001,
      "loss": 0.1617,
      "step": 2500
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.24532224532224534,
      "eval_f1_macro": 0.6081428044829309,
      "eval_f1_micro": 0.7682759232167399,
      "eval_loss": 0.15029709041118622,
      "eval_roc_auc": 0.8352362538484075,
      "eval_runtime": 434.6587,
      "eval_samples_per_second": 6.64,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 2730
    },
    {
      "epoch": 10.989010989010989,
      "grad_norm": 0.16183075308799744,
      "learning_rate": 0.001,
      "loss": 0.1615,
      "step": 3000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.2609147609147609,
      "eval_f1_macro": 0.6199915554248129,
      "eval_f1_micro": 0.7756647297059341,
      "eval_loss": 0.14407172799110413,
      "eval_roc_auc": 0.8408945796610934,
      "eval_runtime": 431.8807,
      "eval_samples_per_second": 6.682,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.001,
      "step": 3003
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.2494802494802495,
      "eval_f1_macro": 0.6299207659511814,
      "eval_f1_micro": 0.781485559413907,
      "eval_loss": 0.14866559207439423,
      "eval_roc_auc": 0.8542998474050816,
      "eval_runtime": 435.7397,
      "eval_samples_per_second": 6.623,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 3276
    },
    {
      "epoch": 12.820512820512821,
      "grad_norm": 0.17374463379383087,
      "learning_rate": 0.001,
      "loss": 0.1614,
      "step": 3500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.25190575190575193,
      "eval_f1_macro": 0.6241659824597257,
      "eval_f1_micro": 0.7779037321241716,
      "eval_loss": 0.14902691543102264,
      "eval_roc_auc": 0.8445867301441496,
      "eval_runtime": 444.1918,
      "eval_samples_per_second": 6.497,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.001,
      "step": 3549
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.26056826056826055,
      "eval_f1_macro": 0.6378982802249643,
      "eval_f1_micro": 0.7826389795829524,
      "eval_loss": 0.14337006211280823,
      "eval_roc_auc": 0.8474976901507599,
      "eval_runtime": 437.9843,
      "eval_samples_per_second": 6.589,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.001,
      "step": 3822
    },
    {
      "epoch": 14.652014652014651,
      "grad_norm": 0.18867848813533783,
      "learning_rate": 0.001,
      "loss": 0.1599,
      "step": 4000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.2553707553707554,
      "eval_f1_macro": 0.639716503598517,
      "eval_f1_micro": 0.7873585308562887,
      "eval_loss": 0.14354591071605682,
      "eval_roc_auc": 0.8551790656297652,
      "eval_runtime": 440.9279,
      "eval_samples_per_second": 6.545,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 4095
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.25675675675675674,
      "eval_f1_macro": 0.6343613127126344,
      "eval_f1_micro": 0.7792974686292388,
      "eval_loss": 0.1439499706029892,
      "eval_roc_auc": 0.8464149986210657,
      "eval_runtime": 436.9638,
      "eval_samples_per_second": 6.605,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.001,
      "step": 4368
    },
    {
      "epoch": 16.483516483516482,
      "grad_norm": 0.16403253376483917,
      "learning_rate": 0.001,
      "loss": 0.1589,
      "step": 4500
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.2543312543312543,
      "eval_f1_macro": 0.6422270697798029,
      "eval_f1_micro": 0.787784461363732,
      "eval_loss": 0.14478015899658203,
      "eval_roc_auc": 0.8595889192695618,
      "eval_runtime": 441.556,
      "eval_samples_per_second": 6.536,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 4641
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.25675675675675674,
      "eval_f1_macro": 0.6417123667888478,
      "eval_f1_micro": 0.786493860845839,
      "eval_loss": 0.14397625625133514,
      "eval_roc_auc": 0.8551892735001003,
      "eval_runtime": 445.1896,
      "eval_samples_per_second": 6.483,
      "eval_steps_per_second": 0.204,
      "learning_rate": 0.001,
      "step": 4914
    },
    {
      "epoch": 18.315018315018314,
      "grad_norm": 1.3815889358520508,
      "learning_rate": 0.001,
      "loss": 0.1604,
      "step": 5000
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.253984753984754,
      "eval_f1_macro": 0.6317583185615991,
      "eval_f1_micro": 0.7863510343356792,
      "eval_loss": 0.14199253916740417,
      "eval_roc_auc": 0.854983895441361,
      "eval_runtime": 435.0664,
      "eval_samples_per_second": 6.633,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 5187
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.2588357588357588,
      "eval_f1_macro": 0.6408966299078661,
      "eval_f1_micro": 0.7868513006341401,
      "eval_loss": 0.14092272520065308,
      "eval_roc_auc": 0.8521780322337986,
      "eval_runtime": 434.5134,
      "eval_samples_per_second": 6.642,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 5460
    },
    {
      "epoch": 20.146520146520146,
      "grad_norm": 0.21049675345420837,
      "learning_rate": 0.001,
      "loss": 0.1586,
      "step": 5500
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.26195426195426197,
      "eval_f1_macro": 0.6412583916380257,
      "eval_f1_micro": 0.7864882090503504,
      "eval_loss": 0.1425119787454605,
      "eval_roc_auc": 0.8561061843865996,
      "eval_runtime": 433.1377,
      "eval_samples_per_second": 6.663,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.001,
      "step": 5733
    },
    {
      "epoch": 21.978021978021978,
      "grad_norm": 0.159688800573349,
      "learning_rate": 0.001,
      "loss": 0.1587,
      "step": 6000
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.23700623700623702,
      "eval_f1_macro": 0.6371452798177432,
      "eval_f1_micro": 0.7854284761587195,
      "eval_loss": 0.15379400551319122,
      "eval_roc_auc": 0.860841268018702,
      "eval_runtime": 439.1581,
      "eval_samples_per_second": 6.572,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 6006
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.25571725571725573,
      "eval_f1_macro": 0.6390434486158698,
      "eval_f1_micro": 0.7841676771176165,
      "eval_loss": 0.1418805718421936,
      "eval_roc_auc": 0.8497106920533675,
      "eval_runtime": 434.4927,
      "eval_samples_per_second": 6.642,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 6279
    },
    {
      "epoch": 23.80952380952381,
      "grad_norm": 0.20623169839382172,
      "learning_rate": 0.001,
      "loss": 0.1592,
      "step": 6500
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.2598752598752599,
      "eval_f1_macro": 0.6458978920546691,
      "eval_f1_micro": 0.7869535635312129,
      "eval_loss": 0.14135514199733734,
      "eval_roc_auc": 0.8561374786855376,
      "eval_runtime": 437.945,
      "eval_samples_per_second": 6.59,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.001,
      "step": 6552
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.26853776853776856,
      "eval_f1_macro": 0.6262981090846956,
      "eval_f1_micro": 0.786773581652009,
      "eval_loss": 0.13985148072242737,
      "eval_roc_auc": 0.8523112308572252,
      "eval_runtime": 433.5621,
      "eval_samples_per_second": 6.656,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.001,
      "step": 6825
    },
    {
      "epoch": 25.641025641025642,
      "grad_norm": 0.167380690574646,
      "learning_rate": 0.001,
      "loss": 0.1586,
      "step": 7000
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.2591822591822592,
      "eval_f1_macro": 0.6237830069375186,
      "eval_f1_micro": 0.7846557710221018,
      "eval_loss": 0.14649754762649536,
      "eval_roc_auc": 0.8560739377107973,
      "eval_runtime": 436.7443,
      "eval_samples_per_second": 6.608,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.001,
      "step": 7098
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.23804573804573806,
      "eval_f1_macro": 0.6344307952131357,
      "eval_f1_micro": 0.7719951506754418,
      "eval_loss": 0.15506784617900848,
      "eval_roc_auc": 0.8432688820115058,
      "eval_runtime": 431.1317,
      "eval_samples_per_second": 6.694,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.001,
      "step": 7371
    },
    {
      "epoch": 27.47252747252747,
      "grad_norm": 0.17562341690063477,
      "learning_rate": 0.001,
      "loss": 0.16,
      "step": 7500
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.2616077616077616,
      "eval_f1_macro": 0.6429949936408241,
      "eval_f1_micro": 0.7891238152420981,
      "eval_loss": 0.14431345462799072,
      "eval_roc_auc": 0.8549858001950897,
      "eval_runtime": 430.551,
      "eval_samples_per_second": 6.703,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.001,
      "step": 7644
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.25675675675675674,
      "eval_f1_macro": 0.6415824285032449,
      "eval_f1_micro": 0.7873995663818392,
      "eval_loss": 0.14275498688220978,
      "eval_roc_auc": 0.8564611929231155,
      "eval_runtime": 443.3199,
      "eval_samples_per_second": 6.51,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.001,
      "step": 7917
    },
    {
      "epoch": 29.304029304029303,
      "grad_norm": 0.16407011449337006,
      "learning_rate": 0.001,
      "loss": 0.1589,
      "step": 8000
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.2525987525987526,
      "eval_f1_macro": 0.6308133523221491,
      "eval_f1_micro": 0.7798808735936467,
      "eval_loss": 0.14164045453071594,
      "eval_roc_auc": 0.8425187578001007,
      "eval_runtime": 435.3871,
      "eval_samples_per_second": 6.629,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.001,
      "step": 8190
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.26888426888426886,
      "eval_f1_macro": 0.6431010910213645,
      "eval_f1_micro": 0.7895365707945718,
      "eval_loss": 0.13976627588272095,
      "eval_roc_auc": 0.8565786303801245,
      "eval_runtime": 429.7695,
      "eval_samples_per_second": 6.715,
      "eval_steps_per_second": 0.212,
      "learning_rate": 0.001,
      "step": 8463
    },
    {
      "epoch": 31.135531135531135,
      "grad_norm": 0.3522001802921295,
      "learning_rate": 0.001,
      "loss": 0.1588,
      "step": 8500
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.25675675675675674,
      "eval_f1_macro": 0.6520927708015384,
      "eval_f1_micro": 0.7891036166898235,
      "eval_loss": 0.1448184847831726,
      "eval_roc_auc": 0.8600551412790717,
      "eval_runtime": 437.4042,
      "eval_samples_per_second": 6.598,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.001,
      "step": 8736
    },
    {
      "epoch": 32.967032967032964,
      "grad_norm": 0.16505810618400574,
      "learning_rate": 0.001,
      "loss": 0.1581,
      "step": 9000
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.26403326403326405,
      "eval_f1_macro": 0.6496848321151188,
      "eval_f1_micro": 0.7895652173913044,
      "eval_loss": 0.14042973518371582,
      "eval_roc_auc": 0.8582461081320644,
      "eval_runtime": 438.9972,
      "eval_samples_per_second": 6.574,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 9009
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.25571725571725573,
      "eval_f1_macro": 0.6448790211155284,
      "eval_f1_micro": 0.7870906828033133,
      "eval_loss": 0.1426127403974533,
      "eval_roc_auc": 0.8537051886564672,
      "eval_runtime": 443.0007,
      "eval_samples_per_second": 6.515,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.001,
      "step": 9282
    },
    {
      "epoch": 34.798534798534796,
      "grad_norm": 0.17374606430530548,
      "learning_rate": 0.001,
      "loss": 0.1578,
      "step": 9500
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.262993762993763,
      "eval_f1_macro": 0.6428423378015612,
      "eval_f1_micro": 0.7846327880264532,
      "eval_loss": 0.14135821163654327,
      "eval_roc_auc": 0.8487055052211715,
      "eval_runtime": 439.0163,
      "eval_samples_per_second": 6.574,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 9555
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.26784476784476785,
      "eval_f1_macro": 0.6434020884943297,
      "eval_f1_micro": 0.7834209497328063,
      "eval_loss": 0.14652539789676666,
      "eval_roc_auc": 0.848444661270401,
      "eval_runtime": 441.3143,
      "eval_samples_per_second": 6.54,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 9828
    },
    {
      "epoch": 36.63003663003663,
      "grad_norm": 0.1908567249774933,
      "learning_rate": 0.001,
      "loss": 0.1576,
      "step": 10000
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.2668052668052668,
      "eval_f1_macro": 0.6438477431550106,
      "eval_f1_micro": 0.792425408224331,
      "eval_loss": 0.13795886933803558,
      "eval_roc_auc": 0.8576696786814598,
      "eval_runtime": 438.7851,
      "eval_samples_per_second": 6.577,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 10101
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.2636867636867637,
      "eval_f1_macro": 0.6475331965590188,
      "eval_f1_micro": 0.7892280686732029,
      "eval_loss": 0.13921019434928894,
      "eval_roc_auc": 0.8555102379171546,
      "eval_runtime": 440.2794,
      "eval_samples_per_second": 6.555,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.001,
      "step": 10374
    },
    {
      "epoch": 38.46153846153846,
      "grad_norm": 0.17312012612819672,
      "learning_rate": 0.001,
      "loss": 0.1556,
      "step": 10500
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.24601524601524602,
      "eval_f1_macro": 0.659217552215385,
      "eval_f1_micro": 0.7871620243872598,
      "eval_loss": 0.14584119617938995,
      "eval_roc_auc": 0.8679696612972285,
      "eval_runtime": 440.7666,
      "eval_samples_per_second": 6.548,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 10647
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.26992376992376993,
      "eval_f1_macro": 0.6469476365862663,
      "eval_f1_micro": 0.79463243873979,
      "eval_loss": 0.1389026641845703,
      "eval_roc_auc": 0.8659848006017948,
      "eval_runtime": 449.7344,
      "eval_samples_per_second": 6.417,
      "eval_steps_per_second": 0.202,
      "learning_rate": 0.001,
      "step": 10920
    },
    {
      "epoch": 40.29304029304029,
      "grad_norm": 0.17165251076221466,
      "learning_rate": 0.001,
      "loss": 0.1577,
      "step": 11000
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.2616077616077616,
      "eval_f1_macro": 0.6509894683187031,
      "eval_f1_micro": 0.784842032071618,
      "eval_loss": 0.14020991325378418,
      "eval_roc_auc": 0.8491298897174419,
      "eval_runtime": 442.6364,
      "eval_samples_per_second": 6.52,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.001,
      "step": 11193
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.27165627165627165,
      "eval_f1_macro": 0.6608924914997423,
      "eval_f1_micro": 0.7927685516081564,
      "eval_loss": 0.14042720198631287,
      "eval_roc_auc": 0.8624648827798459,
      "eval_runtime": 447.1222,
      "eval_samples_per_second": 6.455,
      "eval_steps_per_second": 0.204,
      "learning_rate": 0.001,
      "step": 11466
    },
    {
      "epoch": 42.124542124542124,
      "grad_norm": 0.16945631802082062,
      "learning_rate": 0.001,
      "loss": 0.1576,
      "step": 11500
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.2695772695772696,
      "eval_f1_macro": 0.6427022769326964,
      "eval_f1_micro": 0.7930726352070125,
      "eval_loss": 0.13943640887737274,
      "eval_roc_auc": 0.8592814088524369,
      "eval_runtime": 443.0609,
      "eval_samples_per_second": 6.514,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.001,
      "step": 11739
    },
    {
      "epoch": 43.956043956043956,
      "grad_norm": 0.17723415791988373,
      "learning_rate": 0.0001,
      "loss": 0.1543,
      "step": 12000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.27546777546777546,
      "eval_f1_macro": 0.6567716426576066,
      "eval_f1_micro": 0.7989137353078458,
      "eval_loss": 0.1367315948009491,
      "eval_roc_auc": 0.8632369250728903,
      "eval_runtime": 444.9892,
      "eval_samples_per_second": 6.486,
      "eval_steps_per_second": 0.204,
      "learning_rate": 0.0001,
      "step": 12012
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.28274428274428276,
      "eval_f1_macro": 0.6686203083248894,
      "eval_f1_micro": 0.8018308187828446,
      "eval_loss": 0.13616175949573517,
      "eval_roc_auc": 0.8651714707596159,
      "eval_runtime": 435.6671,
      "eval_samples_per_second": 6.624,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 12285
    },
    {
      "epoch": 45.78754578754579,
      "grad_norm": 0.15646368265151978,
      "learning_rate": 0.0001,
      "loss": 0.1481,
      "step": 12500
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.2851697851697852,
      "eval_f1_macro": 0.6640104860714046,
      "eval_f1_micro": 0.8021852369457503,
      "eval_loss": 0.13375289738178253,
      "eval_roc_auc": 0.8655685984983028,
      "eval_runtime": 439.641,
      "eval_samples_per_second": 6.564,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 12558
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.2785862785862786,
      "eval_f1_macro": 0.65726703563479,
      "eval_f1_micro": 0.7998804746862461,
      "eval_loss": 0.14095526933670044,
      "eval_roc_auc": 0.8620771714997334,
      "eval_runtime": 432.1404,
      "eval_samples_per_second": 6.678,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.0001,
      "step": 12831
    },
    {
      "epoch": 47.61904761904762,
      "grad_norm": 0.14561912417411804,
      "learning_rate": 0.0001,
      "loss": 0.1472,
      "step": 13000
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.28482328482328484,
      "eval_f1_macro": 0.6728387979723557,
      "eval_f1_micro": 0.8044442566853957,
      "eval_loss": 0.13375185430049896,
      "eval_roc_auc": 0.8674991066436737,
      "eval_runtime": 444.9787,
      "eval_samples_per_second": 6.486,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 13104
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.2855162855162855,
      "eval_f1_macro": 0.674164075762875,
      "eval_f1_micro": 0.8058309037900874,
      "eval_loss": 0.13221527636051178,
      "eval_roc_auc": 0.8723556652741397,
      "eval_runtime": 443.2861,
      "eval_samples_per_second": 6.51,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 13377
    },
    {
      "epoch": 49.45054945054945,
      "grad_norm": 0.18880312144756317,
      "learning_rate": 0.0001,
      "loss": 0.1448,
      "step": 13500
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.28967428967428965,
      "eval_f1_macro": 0.6738599949249782,
      "eval_f1_micro": 0.8062985513331933,
      "eval_loss": 0.13315953314304352,
      "eval_roc_auc": 0.8702548292213903,
      "eval_runtime": 434.9136,
      "eval_samples_per_second": 6.636,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 13650
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.28967428967428965,
      "eval_f1_macro": 0.6770873238469556,
      "eval_f1_micro": 0.8062836021505377,
      "eval_loss": 0.13057135045528412,
      "eval_roc_auc": 0.8701618987014408,
      "eval_runtime": 431.3628,
      "eval_samples_per_second": 6.69,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.0001,
      "step": 13923
    },
    {
      "epoch": 51.282051282051285,
      "grad_norm": 0.17863284051418304,
      "learning_rate": 0.0001,
      "loss": 0.1432,
      "step": 14000
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.2872487872487873,
      "eval_f1_macro": 0.6726562275384118,
      "eval_f1_micro": 0.8043922369765066,
      "eval_loss": 0.13108478486537933,
      "eval_roc_auc": 0.8653604706190395,
      "eval_runtime": 432.9945,
      "eval_samples_per_second": 6.665,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.0001,
      "step": 14196
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.2872487872487873,
      "eval_f1_macro": 0.6702824874792834,
      "eval_f1_micro": 0.8070734160241367,
      "eval_loss": 0.13161474466323853,
      "eval_roc_auc": 0.8712745968092424,
      "eval_runtime": 437.0691,
      "eval_samples_per_second": 6.603,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.0001,
      "step": 14469
    },
    {
      "epoch": 53.11355311355312,
      "grad_norm": 0.15456052124500275,
      "learning_rate": 0.0001,
      "loss": 0.1438,
      "step": 14500
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.2882882882882883,
      "eval_f1_macro": 0.6787531928667037,
      "eval_f1_micro": 0.8064162093710426,
      "eval_loss": 0.1315840184688568,
      "eval_roc_auc": 0.8688063579069815,
      "eval_runtime": 432.9273,
      "eval_samples_per_second": 6.666,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.0001,
      "step": 14742
    },
    {
      "epoch": 54.94505494505494,
      "grad_norm": 0.20175035297870636,
      "learning_rate": 0.0001,
      "loss": 0.1417,
      "step": 15000
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.2875952875952876,
      "eval_f1_macro": 0.6698514928377199,
      "eval_f1_micro": 0.8061478697800111,
      "eval_loss": 0.13084293901920319,
      "eval_roc_auc": 0.8685749571564548,
      "eval_runtime": 434.5176,
      "eval_samples_per_second": 6.642,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 15015
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.29417879417879417,
      "eval_f1_macro": 0.6799502024965028,
      "eval_f1_micro": 0.8094286190238215,
      "eval_loss": 0.12969879806041718,
      "eval_roc_auc": 0.8743735506433774,
      "eval_runtime": 432.6673,
      "eval_samples_per_second": 6.67,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.0001,
      "step": 15288
    },
    {
      "epoch": 56.776556776556774,
      "grad_norm": 0.2052290290594101,
      "learning_rate": 0.0001,
      "loss": 0.1415,
      "step": 15500
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.2934857934857935,
      "eval_f1_macro": 0.6716759101412201,
      "eval_f1_micro": 0.8086806577785254,
      "eval_loss": 0.1296372264623642,
      "eval_roc_auc": 0.8711020824592034,
      "eval_runtime": 430.2029,
      "eval_samples_per_second": 6.708,
      "eval_steps_per_second": 0.212,
      "learning_rate": 0.0001,
      "step": 15561
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.29244629244629244,
      "eval_f1_macro": 0.6784509633805341,
      "eval_f1_micro": 0.8068982880161129,
      "eval_loss": 0.12973745167255402,
      "eval_roc_auc": 0.8707904271906546,
      "eval_runtime": 437.8438,
      "eval_samples_per_second": 6.591,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.0001,
      "step": 15834
    },
    {
      "epoch": 58.608058608058606,
      "grad_norm": 0.18146342039108276,
      "learning_rate": 0.0001,
      "loss": 0.1413,
      "step": 16000
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.2910602910602911,
      "eval_f1_macro": 0.6811347101829983,
      "eval_f1_micro": 0.8087436297013858,
      "eval_loss": 0.12995606660842896,
      "eval_roc_auc": 0.8707232376735605,
      "eval_runtime": 440.1545,
      "eval_samples_per_second": 6.557,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 16107
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.28794178794178793,
      "eval_f1_macro": 0.6725887638706813,
      "eval_f1_micro": 0.8056052474657126,
      "eval_loss": 0.13024823367595673,
      "eval_roc_auc": 0.8658006699367622,
      "eval_runtime": 442.8738,
      "eval_samples_per_second": 6.517,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 16380
    },
    {
      "epoch": 60.43956043956044,
      "grad_norm": 0.18909117579460144,
      "learning_rate": 0.0001,
      "loss": 0.1404,
      "step": 16500
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.2948717948717949,
      "eval_f1_macro": 0.6842961167409227,
      "eval_f1_micro": 0.8095537925534148,
      "eval_loss": 0.12872998416423798,
      "eval_roc_auc": 0.8721413670658958,
      "eval_runtime": 442.546,
      "eval_samples_per_second": 6.521,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 16653
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.29002079002079,
      "eval_f1_macro": 0.6821531683206365,
      "eval_f1_micro": 0.8079526226734349,
      "eval_loss": 0.12909561395645142,
      "eval_roc_auc": 0.8690393280672706,
      "eval_runtime": 439.5646,
      "eval_samples_per_second": 6.566,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 16926
    },
    {
      "epoch": 62.27106227106227,
      "grad_norm": 0.18229062855243683,
      "learning_rate": 0.0001,
      "loss": 0.1393,
      "step": 17000
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.29799029799029797,
      "eval_f1_macro": 0.6812919501021206,
      "eval_f1_micro": 0.8075538806791719,
      "eval_loss": 0.12872986495494843,
      "eval_roc_auc": 0.8685427641356871,
      "eval_runtime": 439.5217,
      "eval_samples_per_second": 6.566,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 17199
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.2959112959112959,
      "eval_f1_macro": 0.6805602232602442,
      "eval_f1_micro": 0.8090726144558109,
      "eval_loss": 0.12864243984222412,
      "eval_roc_auc": 0.8722296430115927,
      "eval_runtime": 436.4112,
      "eval_samples_per_second": 6.613,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 17472
    },
    {
      "epoch": 64.1025641025641,
      "grad_norm": 0.2255202978849411,
      "learning_rate": 0.0001,
      "loss": 0.1395,
      "step": 17500
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.29313929313929316,
      "eval_f1_macro": 0.6837997472607307,
      "eval_f1_micro": 0.809268560334276,
      "eval_loss": 0.12800218164920807,
      "eval_roc_auc": 0.8704023269115411,
      "eval_runtime": 436.6673,
      "eval_samples_per_second": 6.609,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.0001,
      "step": 17745
    },
    {
      "epoch": 65.93406593406593,
      "grad_norm": 0.20691530406475067,
      "learning_rate": 0.0001,
      "loss": 0.1389,
      "step": 18000
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.2959112959112959,
      "eval_f1_macro": 0.685457875933014,
      "eval_f1_micro": 0.8107521495951249,
      "eval_loss": 0.12777170538902283,
      "eval_roc_auc": 0.8744479165213509,
      "eval_runtime": 443.2239,
      "eval_samples_per_second": 6.511,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 18018
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.2948717948717949,
      "eval_f1_macro": 0.6849396578990685,
      "eval_f1_micro": 0.8098450774612694,
      "eval_loss": 0.12816764414310455,
      "eval_roc_auc": 0.874604334361006,
      "eval_runtime": 435.2924,
      "eval_samples_per_second": 6.63,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 18291
    },
    {
      "epoch": 67.76556776556777,
      "grad_norm": 0.23671405017375946,
      "learning_rate": 0.0001,
      "loss": 0.1376,
      "step": 18500
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.29799029799029797,
      "eval_f1_macro": 0.6903099963278952,
      "eval_f1_micro": 0.8123470107455503,
      "eval_loss": 0.12804801762104034,
      "eval_roc_auc": 0.8771048199390693,
      "eval_runtime": 435.6555,
      "eval_samples_per_second": 6.625,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 18564
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.6800351861453543,
      "eval_f1_micro": 0.8104663431103608,
      "eval_loss": 0.12803924083709717,
      "eval_roc_auc": 0.8710816667185332,
      "eval_runtime": 435.4705,
      "eval_samples_per_second": 6.627,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 18837
    },
    {
      "epoch": 69.59706959706959,
      "grad_norm": 0.19641809165477753,
      "learning_rate": 0.0001,
      "loss": 0.1375,
      "step": 19000
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.29313929313929316,
      "eval_f1_macro": 0.684802818649885,
      "eval_f1_micro": 0.8096462751380749,
      "eval_loss": 0.12764029204845428,
      "eval_roc_auc": 0.8709255305678097,
      "eval_runtime": 438.0603,
      "eval_samples_per_second": 6.588,
      "eval_steps_per_second": 0.208,
      "learning_rate": 0.0001,
      "step": 19110
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.29036729036729036,
      "eval_f1_macro": 0.6796736257485385,
      "eval_f1_micro": 0.8072724183339705,
      "eval_loss": 0.12794704735279083,
      "eval_roc_auc": 0.8674678659997703,
      "eval_runtime": 438.6839,
      "eval_samples_per_second": 6.579,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 19383
    },
    {
      "epoch": 71.42857142857143,
      "grad_norm": 0.23725061118602753,
      "learning_rate": 0.0001,
      "loss": 0.1368,
      "step": 19500
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.2938322938322938,
      "eval_f1_macro": 0.6802343842914587,
      "eval_f1_micro": 0.8102650399663442,
      "eval_loss": 0.12780210375785828,
      "eval_roc_auc": 0.8718861300662425,
      "eval_runtime": 436.1264,
      "eval_samples_per_second": 6.617,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 19656
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.29764379764379767,
      "eval_f1_macro": 0.6805723882610378,
      "eval_f1_micro": 0.8091473263623224,
      "eval_loss": 0.12723641097545624,
      "eval_roc_auc": 0.8683039589903839,
      "eval_runtime": 430.6869,
      "eval_samples_per_second": 6.701,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.0001,
      "step": 19929
    },
    {
      "epoch": 73.26007326007326,
      "grad_norm": 0.2105712741613388,
      "learning_rate": 0.0001,
      "loss": 0.137,
      "step": 20000
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.2934857934857935,
      "eval_f1_macro": 0.6777188921642516,
      "eval_f1_micro": 0.8064391831142698,
      "eval_loss": 0.12804573774337769,
      "eval_roc_auc": 0.8647877433658233,
      "eval_runtime": 435.4406,
      "eval_samples_per_second": 6.628,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 20202
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.29244629244629244,
      "eval_f1_macro": 0.6885203936930924,
      "eval_f1_micro": 0.8109922383050138,
      "eval_loss": 0.1273234635591507,
      "eval_roc_auc": 0.8730787473480999,
      "eval_runtime": 433.0115,
      "eval_samples_per_second": 6.665,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.0001,
      "step": 20475
    },
    {
      "epoch": 75.0915750915751,
      "grad_norm": 0.24889850616455078,
      "learning_rate": 0.0001,
      "loss": 0.1367,
      "step": 20500
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.2972972972972973,
      "eval_f1_macro": 0.6810578369044884,
      "eval_f1_micro": 0.8088975345709815,
      "eval_loss": 0.1272992193698883,
      "eval_roc_auc": 0.8696399827660029,
      "eval_runtime": 432.2403,
      "eval_samples_per_second": 6.677,
      "eval_steps_per_second": 0.211,
      "learning_rate": 0.0001,
      "step": 20748
    },
    {
      "epoch": 76.92307692307692,
      "grad_norm": 0.24539624154567719,
      "learning_rate": 0.0001,
      "loss": 0.1358,
      "step": 21000
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.29244629244629244,
      "eval_f1_macro": 0.6863183190306963,
      "eval_f1_micro": 0.8102101349375445,
      "eval_loss": 0.12745273113250732,
      "eval_roc_auc": 0.8738921845003481,
      "eval_runtime": 434.2537,
      "eval_samples_per_second": 6.646,
      "eval_steps_per_second": 0.21,
      "learning_rate": 0.0001,
      "step": 21021
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.2945252945252945,
      "eval_f1_macro": 0.6897104532016692,
      "eval_f1_micro": 0.8121675531914894,
      "eval_loss": 0.12705788016319275,
      "eval_roc_auc": 0.876527719800532,
      "eval_runtime": 444.7222,
      "eval_samples_per_second": 6.489,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 21294
    },
    {
      "epoch": 78.75457875457876,
      "grad_norm": 0.23895101249217987,
      "learning_rate": 0.0001,
      "loss": 0.1352,
      "step": 21500
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.2934857934857935,
      "eval_f1_macro": 0.6881838490414868,
      "eval_f1_micro": 0.809842452990005,
      "eval_loss": 0.12710121273994446,
      "eval_roc_auc": 0.869729211088564,
      "eval_runtime": 435.3871,
      "eval_samples_per_second": 6.629,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 21567
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.2983367983367983,
      "eval_f1_macro": 0.6914032136958002,
      "eval_f1_micro": 0.8123911420751431,
      "eval_loss": 0.12715762853622437,
      "eval_roc_auc": 0.8772640412542829,
      "eval_runtime": 435.2387,
      "eval_samples_per_second": 6.631,
      "eval_steps_per_second": 0.209,
      "learning_rate": 0.0001,
      "step": 21840
    },
    {
      "epoch": 80.58608058608058,
      "grad_norm": 0.28309133648872375,
      "learning_rate": 0.0001,
      "loss": 0.1353,
      "step": 22000
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.2966042966042966,
      "eval_f1_macro": 0.6899389708752343,
      "eval_f1_micro": 0.810378232667846,
      "eval_loss": 0.12650521099567413,
      "eval_roc_auc": 0.8716067214734959,
      "eval_runtime": 440.2197,
      "eval_samples_per_second": 6.556,
      "eval_steps_per_second": 0.207,
      "learning_rate": 0.0001,
      "step": 22113
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.29140679140679143,
      "eval_f1_macro": 0.6844864031747653,
      "eval_f1_micro": 0.8105446364138047,
      "eval_loss": 0.12635371088981628,
      "eval_roc_auc": 0.8694408286912787,
      "eval_runtime": 441.9602,
      "eval_samples_per_second": 6.53,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 22386
    },
    {
      "epoch": 82.41758241758242,
      "grad_norm": 0.24017925560474396,
      "learning_rate": 0.0001,
      "loss": 0.1337,
      "step": 22500
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.2934857934857935,
      "eval_f1_macro": 0.6832392344549459,
      "eval_f1_micro": 0.8099670022844573,
      "eval_loss": 0.1272997260093689,
      "eval_roc_auc": 0.8701139445142557,
      "eval_runtime": 442.0471,
      "eval_samples_per_second": 6.529,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 22659
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.2959112959112959,
      "eval_f1_macro": 0.6944491344986764,
      "eval_f1_micro": 0.8124478558318038,
      "eval_loss": 0.12640425562858582,
      "eval_roc_auc": 0.875556912431764,
      "eval_runtime": 443.3633,
      "eval_samples_per_second": 6.509,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 22932
    },
    {
      "epoch": 84.24908424908425,
      "grad_norm": 0.22998856008052826,
      "learning_rate": 0.0001,
      "loss": 0.1354,
      "step": 23000
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.2972972972972973,
      "eval_f1_macro": 0.6879519222426981,
      "eval_f1_micro": 0.812659392115055,
      "eval_loss": 0.12647400796413422,
      "eval_roc_auc": 0.8750247349709565,
      "eval_runtime": 442.8099,
      "eval_samples_per_second": 6.517,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 23205
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.6933253774763921,
      "eval_f1_micro": 0.8135877542461731,
      "eval_loss": 0.12585221230983734,
      "eval_roc_auc": 0.8746291766420319,
      "eval_runtime": 440.9089,
      "eval_samples_per_second": 6.546,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 23478
    },
    {
      "epoch": 86.08058608058609,
      "grad_norm": 0.29592010378837585,
      "learning_rate": 0.0001,
      "loss": 0.1334,
      "step": 23500
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.2966042966042966,
      "eval_f1_macro": 0.6882459007361815,
      "eval_f1_micro": 0.8111366966715512,
      "eval_loss": 0.12641744315624237,
      "eval_roc_auc": 0.8737689120583252,
      "eval_runtime": 444.5125,
      "eval_samples_per_second": 6.493,
      "eval_steps_per_second": 0.205,
      "learning_rate": 0.0001,
      "step": 23751
    },
    {
      "epoch": 87.91208791208791,
      "grad_norm": 0.252650648355484,
      "learning_rate": 0.0001,
      "loss": 0.1335,
      "step": 24000
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.6859575429209334,
      "eval_f1_micro": 0.8126931106471816,
      "eval_loss": 0.1263686865568161,
      "eval_roc_auc": 0.8754230967754396,
      "eval_runtime": 448.2495,
      "eval_samples_per_second": 6.438,
      "eval_steps_per_second": 0.203,
      "learning_rate": 0.0001,
      "step": 24024
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.2983367983367983,
      "eval_f1_macro": 0.6990366097632199,
      "eval_f1_micro": 0.8140188460902628,
      "eval_loss": 0.12690132856369019,
      "eval_roc_auc": 0.8791768834795075,
      "eval_runtime": 441.4136,
      "eval_samples_per_second": 6.538,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 24297
    },
    {
      "epoch": 89.74358974358974,
      "grad_norm": 0.2610660791397095,
      "learning_rate": 0.0001,
      "loss": 0.1332,
      "step": 24500
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.29799029799029797,
      "eval_f1_macro": 0.6994167448254883,
      "eval_f1_micro": 0.8155163144617673,
      "eval_loss": 0.12612390518188477,
      "eval_roc_auc": 0.8798283494960182,
      "eval_runtime": 445.6794,
      "eval_samples_per_second": 6.476,
      "eval_steps_per_second": 0.204,
      "learning_rate": 0.0001,
      "step": 24570
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.28932778932778935,
      "eval_f1_macro": 0.6827913109763548,
      "eval_f1_micro": 0.8108811552831535,
      "eval_loss": 0.1268243044614792,
      "eval_roc_auc": 0.8728467295646753,
      "eval_runtime": 442.1487,
      "eval_samples_per_second": 6.527,
      "eval_steps_per_second": 0.206,
      "learning_rate": 0.0001,
      "step": 24843
    },
    {
      "epoch": 91.57509157509158,
      "grad_norm": 0.28306326270103455,
      "learning_rate": 0.0001,
      "loss": 0.1326,
      "step": 25000
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.6858483939371968,
      "eval_f1_micro": 0.8123787840458724,
      "eval_loss": 0.12613284587860107,
      "eval_roc_auc": 0.8724090520335794,
      "eval_runtime": 447.7751,
      "eval_samples_per_second": 6.445,
      "eval_steps_per_second": 0.203,
      "learning_rate": 0.0001,
      "step": 25116
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.2966042966042966,
      "eval_f1_macro": 0.6897216822080747,
      "eval_f1_micro": 0.8138213420238991,
      "eval_loss": 0.1258293092250824,
      "eval_roc_auc": 0.8758502847472687,
      "eval_runtime": 442.7745,
      "eval_samples_per_second": 6.518,
      "eval_steps_per_second": 0.206,
      "learning_rate": 1e-05,
      "step": 25389
    },
    {
      "epoch": 93.4065934065934,
      "grad_norm": 0.36196333169937134,
      "learning_rate": 1e-05,
      "loss": 0.132,
      "step": 25500
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.29764379764379767,
      "eval_f1_macro": 0.6940665827082791,
      "eval_f1_micro": 0.8137706015226304,
      "eval_loss": 0.12682591378688812,
      "eval_roc_auc": 0.8754785626674707,
      "eval_runtime": 447.9681,
      "eval_samples_per_second": 6.442,
      "eval_steps_per_second": 0.203,
      "learning_rate": 1e-05,
      "step": 25662
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.2948717948717949,
      "eval_f1_macro": 0.6913394393323408,
      "eval_f1_micro": 0.8133975298304374,
      "eval_loss": 0.1256789118051529,
      "eval_roc_auc": 0.8750151441335194,
      "eval_runtime": 494.4109,
      "eval_samples_per_second": 5.837,
      "eval_steps_per_second": 0.184,
      "learning_rate": 1e-05,
      "step": 25935
    },
    {
      "epoch": 95.23809523809524,
      "grad_norm": 0.28360626101493835,
      "learning_rate": 1e-05,
      "loss": 0.1294,
      "step": 26000
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.29764379764379767,
      "eval_f1_macro": 0.6957055849225957,
      "eval_f1_micro": 0.8147281313996739,
      "eval_loss": 0.12587758898735046,
      "eval_roc_auc": 0.876265078406102,
      "eval_runtime": 456.43,
      "eval_samples_per_second": 6.323,
      "eval_steps_per_second": 0.199,
      "learning_rate": 1e-05,
      "step": 26208
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.2945252945252945,
      "eval_f1_macro": 0.6940781337907567,
      "eval_f1_micro": 0.8126029480086159,
      "eval_loss": 0.1256256103515625,
      "eval_roc_auc": 0.8720295902150387,
      "eval_runtime": 440.8704,
      "eval_samples_per_second": 6.546,
      "eval_steps_per_second": 0.206,
      "learning_rate": 1e-05,
      "step": 26481
    },
    {
      "epoch": 97.06959706959707,
      "grad_norm": 0.29758304357528687,
      "learning_rate": 1e-05,
      "loss": 0.1302,
      "step": 26500
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.2993762993762994,
      "eval_f1_macro": 0.6951390304078455,
      "eval_f1_micro": 0.8158955813276801,
      "eval_loss": 0.1253080666065216,
      "eval_roc_auc": 0.8785118427392398,
      "eval_runtime": 432.3413,
      "eval_samples_per_second": 6.675,
      "eval_steps_per_second": 0.21,
      "learning_rate": 1e-05,
      "step": 26754
    },
    {
      "epoch": 98.9010989010989,
      "grad_norm": 0.3460623621940613,
      "learning_rate": 1e-05,
      "loss": 0.1298,
      "step": 27000
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.2993762993762994,
      "eval_f1_macro": 0.6968244403216463,
      "eval_f1_micro": 0.8141971169963125,
      "eval_loss": 0.12485036998987198,
      "eval_roc_auc": 0.8751979200089282,
      "eval_runtime": 440.8458,
      "eval_samples_per_second": 6.547,
      "eval_steps_per_second": 0.206,
      "learning_rate": 1e-05,
      "step": 27027
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.2972972972972973,
      "eval_f1_macro": 0.693647218520028,
      "eval_f1_micro": 0.8134507606084869,
      "eval_loss": 0.12519583106040955,
      "eval_roc_auc": 0.8731552996687338,
      "eval_runtime": 426.3213,
      "eval_samples_per_second": 6.77,
      "eval_steps_per_second": 0.213,
      "learning_rate": 1e-05,
      "step": 27300
    },
    {
      "epoch": 100.73260073260073,
      "grad_norm": 0.2845664918422699,
      "learning_rate": 1e-05,
      "loss": 0.1304,
      "step": 27500
    },
    {
      "epoch": 101.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.6961023046950545,
      "eval_f1_micro": 0.8148550421923302,
      "eval_loss": 0.12475299090147018,
      "eval_roc_auc": 0.8764704564648416,
      "eval_runtime": 432.286,
      "eval_samples_per_second": 6.676,
      "eval_steps_per_second": 0.211,
      "learning_rate": 1e-05,
      "step": 27573
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.29625779625779625,
      "eval_f1_macro": 0.692743439816851,
      "eval_f1_micro": 0.81366198367965,
      "eval_loss": 0.12659381330013275,
      "eval_roc_auc": 0.8737986906025352,
      "eval_runtime": 429.8933,
      "eval_samples_per_second": 6.713,
      "eval_steps_per_second": 0.212,
      "learning_rate": 1e-05,
      "step": 27846
    },
    {
      "epoch": 102.56410256410257,
      "grad_norm": 0.276334673166275,
      "learning_rate": 1e-05,
      "loss": 0.1287,
      "step": 28000
    },
    {
      "epoch": 103.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.6954353634647259,
      "eval_f1_micro": 0.8146347596496376,
      "eval_loss": 0.124935083091259,
      "eval_roc_auc": 0.8754020640018715,
      "eval_runtime": 433.4496,
      "eval_samples_per_second": 6.658,
      "eval_steps_per_second": 0.21,
      "learning_rate": 1e-05,
      "step": 28119
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.29764379764379767,
      "eval_f1_macro": 0.692659001716947,
      "eval_f1_micro": 0.8148796863922599,
      "eval_loss": 0.12519653141498566,
      "eval_roc_auc": 0.8769883885139594,
      "eval_runtime": 435.0406,
      "eval_samples_per_second": 6.634,
      "eval_steps_per_second": 0.209,
      "learning_rate": 1e-05,
      "step": 28392
    },
    {
      "epoch": 104.3956043956044,
      "grad_norm": 0.2987622320652008,
      "learning_rate": 1e-05,
      "loss": 0.1282,
      "step": 28500
    },
    {
      "epoch": 105.0,
      "eval_accuracy": 0.29902979902979904,
      "eval_f1_macro": 0.6961790886935857,
      "eval_f1_micro": 0.8152223750573132,
      "eval_loss": 0.12513257563114166,
      "eval_roc_auc": 0.8773257311303291,
      "eval_runtime": 430.7348,
      "eval_samples_per_second": 6.7,
      "eval_steps_per_second": 0.211,
      "learning_rate": 1e-05,
      "step": 28665
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.29972279972279975,
      "eval_f1_macro": 0.6963861386142265,
      "eval_f1_micro": 0.8147252563995664,
      "eval_loss": 0.12511174380779266,
      "eval_roc_auc": 0.8769890055468574,
      "eval_runtime": 434.1489,
      "eval_samples_per_second": 6.647,
      "eval_steps_per_second": 0.21,
      "learning_rate": 1e-05,
      "step": 28938
    },
    {
      "epoch": 106.22710622710623,
      "grad_norm": 0.2862643301486969,
      "learning_rate": 1e-05,
      "loss": 0.1293,
      "step": 29000
    },
    {
      "epoch": 107.0,
      "eval_accuracy": 0.29799029799029797,
      "eval_f1_macro": 0.694620567930595,
      "eval_f1_micro": 0.8144894800685992,
      "eval_loss": 0.12498941272497177,
      "eval_roc_auc": 0.8758874676862554,
      "eval_runtime": 435.6169,
      "eval_samples_per_second": 6.625,
      "eval_steps_per_second": 0.209,
      "learning_rate": 1e-05,
      "step": 29211
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.29972279972279975,
      "eval_f1_macro": 0.6934713387989168,
      "eval_f1_micro": 0.8144792584203683,
      "eval_loss": 0.1248873621225357,
      "eval_roc_auc": 0.8750706100255504,
      "eval_runtime": 426.658,
      "eval_samples_per_second": 6.764,
      "eval_steps_per_second": 0.213,
      "learning_rate": 1.0000000000000002e-06,
      "step": 29484
    },
    {
      "epoch": 108.05860805860806,
      "grad_norm": 0.2752939760684967,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.129,
      "step": 29500
    },
    {
      "epoch": 109.0,
      "eval_accuracy": 0.29521829521829523,
      "eval_f1_macro": 0.6900779361953018,
      "eval_f1_micro": 0.8116150302210575,
      "eval_loss": 0.12527066469192505,
      "eval_roc_auc": 0.8712697142010926,
      "eval_runtime": 434.9611,
      "eval_samples_per_second": 6.635,
      "eval_steps_per_second": 0.209,
      "learning_rate": 1.0000000000000002e-06,
      "step": 29757
    },
    {
      "epoch": 109.89010989010988,
      "grad_norm": 0.2989753484725952,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.1293,
      "step": 30000
    },
    {
      "epoch": 110.0,
      "eval_accuracy": 0.29799029799029797,
      "eval_f1_macro": 0.69491512245201,
      "eval_f1_micro": 0.8143917285082964,
      "eval_loss": 0.125152125954628,
      "eval_roc_auc": 0.8768043785727546,
      "eval_runtime": 434.3393,
      "eval_samples_per_second": 6.645,
      "eval_steps_per_second": 0.21,
      "learning_rate": 1.0000000000000002e-06,
      "step": 30030
    },
    {
      "epoch": 111.0,
      "eval_accuracy": 0.2983367983367983,
      "eval_f1_macro": 0.6932228755688746,
      "eval_f1_micro": 0.8137025263510123,
      "eval_loss": 0.12495684623718262,
      "eval_roc_auc": 0.8754655513215771,
      "eval_runtime": 431.8608,
      "eval_samples_per_second": 6.683,
      "eval_steps_per_second": 0.211,
      "learning_rate": 1.0000000000000002e-06,
      "step": 30303
    },
    {
      "epoch": 111.0,
      "learning_rate": 1.0000000000000002e-06,
      "step": 30303,
      "total_flos": 4.1153080208666034e+20,
      "train_loss": 0.14764341744551096,
      "train_runtime": 198162.6547,
      "train_samples_per_second": 6.598,
      "train_steps_per_second": 0.207
    }
  ],
  "logging_steps": 500,
  "max_steps": 40950,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 150,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 10,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.1153080208666034e+20,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}