diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4864 @@
+{
+  "best_metric": 0.9796821008984106,
+  "best_model_checkpoint": "test/checkpoint-1500",
+  "epoch": 13.377926421404682,
+  "global_step": 4000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "eval_accuracy": 0.8922687113634554,
+      "eval_f1": 0.0,
+      "eval_loss": 1.3138163089752197,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 42.6447,
+      "eval_samples_per_second": 10.06,
+      "eval_steps_per_second": 1.688,
+      "step": 10
+    },
+    {
+      "epoch": 0.07,
+      "eval_accuracy": 0.8922687113634554,
+      "eval_f1": 0.0,
+      "eval_loss": 0.7134402394294739,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 44.6528,
+      "eval_samples_per_second": 9.607,
+      "eval_steps_per_second": 1.612,
+      "step": 20
+    },
+    {
+      "epoch": 0.1,
+      "eval_accuracy": 0.8922687113634554,
+      "eval_f1": 0.0,
+      "eval_loss": 0.6043053865432739,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 44.2765,
+      "eval_samples_per_second": 9.689,
+      "eval_steps_per_second": 1.626,
+      "step": 30
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.8922687113634554,
+      "eval_f1": 0.0,
+      "eval_loss": 0.5121276378631592,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 45.3257,
+      "eval_samples_per_second": 9.465,
+      "eval_steps_per_second": 1.589,
+      "step": 40
+    },
+    {
+      "epoch": 0.17,
+      "eval_accuracy": 0.8922687113634554,
+      "eval_f1": 0.0,
+      "eval_loss": 0.4018501937389374,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 44.7272,
+      "eval_samples_per_second": 9.591,
+      "eval_steps_per_second": 1.61,
+      "step": 50
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.908479477859436,
+      "eval_f1": 0.0,
+      "eval_loss": 0.3501987159252167,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 44.7765,
+      "eval_samples_per_second": 9.581,
+      "eval_steps_per_second": 1.608,
+      "step": 60
+    },
+    {
+      "epoch": 0.23,
+      "eval_accuracy": 0.9237218434108938,
+      "eval_f1": 0.0,
+      "eval_loss": 0.29354768991470337,
+      "eval_precision": 0.0,
+      "eval_recall": 0.0,
+      "eval_runtime": 44.776,
+      "eval_samples_per_second": 9.581,
+      "eval_steps_per_second": 1.608,
+      "step": 70
+    },
+    {
+      "epoch": 0.27,
+      "eval_accuracy": 0.9264413255153751,
+      "eval_f1": 0.0008230452674897119,
+      "eval_loss": 0.24890266358852386,
+      "eval_precision": 0.001594896331738437,
+      "eval_recall": 0.0005546311702717693,
+      "eval_runtime": 44.807,
+      "eval_samples_per_second": 9.574,
+      "eval_steps_per_second": 1.607,
+      "step": 80
+    },
+    {
+      "epoch": 0.3,
+      "eval_accuracy": 0.9392427900559815,
+      "eval_f1": 0.013897161008536827,
+      "eval_loss": 0.2203027307987213,
+      "eval_precision": 0.02445842068483578,
+      "eval_recall": 0.009706045479755962,
+      "eval_runtime": 44.7926,
+      "eval_samples_per_second": 9.577,
+      "eval_steps_per_second": 1.607,
+      "step": 90
+    },
+    {
+      "epoch": 0.33,
+      "eval_accuracy": 0.935727361969701,
+      "eval_f1": 0.00045840018336007336,
+      "eval_loss": 0.19940191507339478,
+      "eval_precision": 0.001321003963011889,
+      "eval_recall": 0.00027731558513588466,
+      "eval_runtime": 44.7253,
+      "eval_samples_per_second": 9.592,
+      "eval_steps_per_second": 1.61,
+      "step": 100
+    },
+    {
+      "epoch": 0.37,
+      "eval_accuracy": 0.943779682152238,
+      "eval_f1": 0.07579030041233065,
+      "eval_loss": 0.1813557893037796,
+      "eval_precision": 0.12979152656355078,
+      "eval_recall": 0.053521907931225736,
+      "eval_runtime": 44.6764,
+      "eval_samples_per_second": 9.602,
+      "eval_steps_per_second": 1.612,
+      "step": 110
+    },
+    {
+      "epoch": 0.4,
+      "eval_accuracy": 0.9479053354912313,
+      "eval_f1": 0.11372837688627163,
+      "eval_loss": 0.16338485479354858,
+      "eval_precision": 0.16903719912472648,
+      "eval_recall": 0.08569051580698835,
+      "eval_runtime": 44.7776,
+      "eval_samples_per_second": 9.581,
+      "eval_steps_per_second": 1.608,
+      "step": 120
+    },
+    {
+      "epoch": 0.43,
+      "eval_accuracy": 0.9628027911172429,
+      "eval_f1": 0.33931930904737473,
+      "eval_loss": 0.14094215631484985,
+      "eval_precision": 0.4426595269968764,
+      "eval_recall": 0.27509706045479754,
+      "eval_runtime": 44.9683,
+      "eval_samples_per_second": 9.54,
+      "eval_steps_per_second": 1.601,
+      "step": 130
+    },
+    {
+      "epoch": 0.47,
+      "eval_accuracy": 0.9663580165026133,
+      "eval_f1": 0.34743875278396436,
+      "eval_loss": 0.12451943010091782,
+      "eval_precision": 0.40746268656716417,
+      "eval_recall": 0.30282861896838603,
+      "eval_runtime": 44.682,
+      "eval_samples_per_second": 9.601,
+      "eval_steps_per_second": 1.611,
+      "step": 140
+    },
+    {
+      "epoch": 0.5,
+      "eval_accuracy": 0.9711734896924995,
+      "eval_f1": 0.47246835443037977,
+      "eval_loss": 0.11197753250598907,
+      "eval_precision": 0.5501105379513633,
+      "eval_recall": 0.4140321686078758,
+      "eval_runtime": 44.7253,
+      "eval_samples_per_second": 9.592,
+      "eval_steps_per_second": 1.61,
+      "step": 150
+    },
+    {
+      "epoch": 0.54,
+      "eval_accuracy": 0.9736674537688043,
+      "eval_f1": 0.5392111368909513,
+      "eval_loss": 0.09976229071617126,
+      "eval_precision": 0.6096537250786989,
+      "eval_recall": 0.48336106489184694,
+      "eval_runtime": 44.7165,
+      "eval_samples_per_second": 9.594,
+      "eval_steps_per_second": 1.61,
+      "step": 160
+    },
+    {
+      "epoch": 0.57,
+      "eval_accuracy": 0.9746623862460534,
+      "eval_f1": 0.5685945368533496,
+      "eval_loss": 0.09186050295829773,
+      "eval_precision": 0.6321683067526298,
+      "eval_recall": 0.5166389351081531,
+      "eval_runtime": 44.7738,
+      "eval_samples_per_second": 9.582,
+      "eval_steps_per_second": 1.608,
+      "step": 170
+    },
+    {
+      "epoch": 0.6,
+      "eval_accuracy": 0.9771696160887214,
+      "eval_f1": 0.5888754534461911,
+      "eval_loss": 0.0845290869474411,
+      "eval_precision": 0.6471760797342193,
+      "eval_recall": 0.5402107598447032,
+      "eval_runtime": 44.7399,
+      "eval_samples_per_second": 9.589,
+      "eval_steps_per_second": 1.609,
+      "step": 180
+    },
+    {
+      "epoch": 0.64,
+      "eval_accuracy": 0.980194210819559,
+      "eval_f1": 0.6471658078205517,
+      "eval_loss": 0.07714465260505676,
+      "eval_precision": 0.7135695187165776,
+      "eval_recall": 0.5920687742651137,
+      "eval_runtime": 44.734,
+      "eval_samples_per_second": 9.59,
+      "eval_steps_per_second": 1.61,
+      "step": 190
+    },
+    {
+      "epoch": 0.67,
+      "eval_accuracy": 0.9795839855668462,
+      "eval_f1": 0.7072773084333779,
+      "eval_loss": 0.07666940242052078,
+      "eval_precision": 0.7596306908627826,
+      "eval_recall": 0.6616749861342207,
+      "eval_runtime": 44.7886,
+      "eval_samples_per_second": 9.578,
+      "eval_steps_per_second": 1.608,
+      "step": 200
+    },
+    {
+      "epoch": 0.7,
+      "eval_accuracy": 0.9832586028494866,
+      "eval_f1": 0.7367480643240022,
+      "eval_loss": 0.06706634908914566,
+      "eval_precision": 0.7954983922829582,
+      "eval_recall": 0.6860787576261785,
+      "eval_runtime": 44.9641,
+      "eval_samples_per_second": 9.541,
+      "eval_steps_per_second": 1.601,
+      "step": 210
+    },
+    {
+      "epoch": 0.74,
+      "eval_accuracy": 0.9847576344485421,
+      "eval_f1": 0.7383826191913097,
+      "eval_loss": 0.06665363162755966,
+      "eval_precision": 0.8097286565188617,
+      "eval_recall": 0.6785912368275097,
+      "eval_runtime": 44.8134,
+      "eval_samples_per_second": 9.573,
+      "eval_steps_per_second": 1.607,
+      "step": 220
+    },
+    {
+      "epoch": 0.77,
+      "eval_accuracy": 0.9845321164203656,
+      "eval_f1": 0.752096513167574,
+      "eval_loss": 0.06071800738573074,
+      "eval_precision": 0.8010028204324663,
+      "eval_recall": 0.7088186356073212,
+      "eval_runtime": 44.7268,
+      "eval_samples_per_second": 9.592,
+      "eval_steps_per_second": 1.61,
+      "step": 230
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.986482184075774,
+      "eval_f1": 0.7756542193472508,
+      "eval_loss": 0.057662855833768845,
+      "eval_precision": 0.8254067584480601,
+      "eval_recall": 0.7315585135884637,
+      "eval_runtime": 44.6791,
+      "eval_samples_per_second": 9.602,
+      "eval_steps_per_second": 1.611,
+      "step": 240
+    },
+    {
+      "epoch": 0.84,
+      "eval_accuracy": 0.9871454723939402,
+      "eval_f1": 0.7851227662356531,
+      "eval_loss": 0.054133981466293335,
+      "eval_precision": 0.8245346353371986,
+      "eval_recall": 0.7493067110371603,
+      "eval_runtime": 44.7641,
+      "eval_samples_per_second": 9.584,
+      "eval_steps_per_second": 1.608,
+      "step": 250
+    },
+    {
+      "epoch": 0.87,
+      "eval_accuracy": 0.9884985805629991,
+      "eval_f1": 0.8150755242704208,
+      "eval_loss": 0.05026474595069885,
+      "eval_precision": 0.8649237472766884,
+      "eval_recall": 0.7706600110926234,
+      "eval_runtime": 44.9736,
+      "eval_samples_per_second": 9.539,
+      "eval_steps_per_second": 1.601,
+      "step": 260
+    },
+    {
+      "epoch": 0.9,
+      "eval_accuracy": 0.9881934679366426,
+      "eval_f1": 0.801503541997976,
+      "eval_loss": 0.04910367354750633,
+      "eval_precision": 0.8372093023255814,
+      "eval_recall": 0.7687188019966722,
+      "eval_runtime": 44.8146,
+      "eval_samples_per_second": 9.573,
+      "eval_steps_per_second": 1.607,
+      "step": 270
+    },
+    {
+      "epoch": 0.94,
+      "eval_accuracy": 0.987875089543923,
+      "eval_f1": 0.8106457242582897,
+      "eval_loss": 0.04733191058039665,
+      "eval_precision": 0.8522935779816514,
+      "eval_recall": 0.7728785357737105,
+      "eval_runtime": 44.8863,
+      "eval_samples_per_second": 9.557,
+      "eval_steps_per_second": 1.604,
+      "step": 280
+    },
+    {
+      "epoch": 0.97,
+      "eval_accuracy": 0.9887904274229923,
+      "eval_f1": 0.8112208387004436,
+      "eval_loss": 0.046026937663555145,
+      "eval_precision": 0.8382135462880804,
+      "eval_recall": 0.785912368275097,
+      "eval_runtime": 44.8674,
+      "eval_samples_per_second": 9.562,
+      "eval_steps_per_second": 1.605,
+      "step": 290
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9894404499747951,
+      "eval_f1": 0.8500071766901106,
+      "eval_loss": 0.0429680198431015,
+      "eval_precision": 0.880987801249628,
+      "eval_recall": 0.8211314475873545,
+      "eval_runtime": 45.1867,
+      "eval_samples_per_second": 9.494,
+      "eval_steps_per_second": 1.593,
+      "step": 300
+    },
+    {
+      "epoch": 1.04,
+      "eval_accuracy": 0.9901170040593245,
+      "eval_f1": 0.8402282453637661,
+      "eval_loss": 0.04167770594358444,
+      "eval_precision": 0.8651586368977673,
+      "eval_recall": 0.8166943982251803,
+      "eval_runtime": 44.8739,
+      "eval_samples_per_second": 9.56,
+      "eval_steps_per_second": 1.604,
+      "step": 310
+    },
+    {
+      "epoch": 1.07,
+      "eval_accuracy": 0.9908333554429439,
+      "eval_f1": 0.8512689635616051,
+      "eval_loss": 0.039857033640146255,
+      "eval_precision": 0.8709022338265158,
+      "eval_recall": 0.8325013865779257,
+      "eval_runtime": 44.8161,
+      "eval_samples_per_second": 9.572,
+      "eval_steps_per_second": 1.607,
+      "step": 320
+    },
+    {
+      "epoch": 1.1,
+      "eval_accuracy": 0.9919211482847364,
+      "eval_f1": 0.8664764621968616,
+      "eval_loss": 0.03763346001505852,
+      "eval_precision": 0.8921856639247944,
+      "eval_recall": 0.8422074320576817,
+      "eval_runtime": 45.0124,
+      "eval_samples_per_second": 9.531,
+      "eval_steps_per_second": 1.6,
+      "step": 330
+    },
+    {
+      "epoch": 1.14,
+      "eval_accuracy": 0.992106869013823,
+      "eval_f1": 0.8503758332151468,
+      "eval_loss": 0.035481277853250504,
+      "eval_precision": 0.8702467343976777,
+      "eval_recall": 0.8313921242373822,
+      "eval_runtime": 44.8357,
+      "eval_samples_per_second": 9.568,
+      "eval_steps_per_second": 1.606,
+      "step": 340
+    },
+    {
+      "epoch": 1.17,
+      "eval_accuracy": 0.9925048420047227,
+      "eval_f1": 0.8676303854875284,
+      "eval_loss": 0.03401191905140877,
+      "eval_precision": 0.8872463768115942,
+      "eval_recall": 0.8488630061009429,
+      "eval_runtime": 44.9137,
+      "eval_samples_per_second": 9.552,
+      "eval_steps_per_second": 1.603,
+      "step": 350
+    },
+    {
+      "epoch": 1.2,
+      "eval_accuracy": 0.9926905627338092,
+      "eval_f1": 0.8718825708992448,
+      "eval_loss": 0.032262638211250305,
+      "eval_precision": 0.8968044561712107,
+      "eval_recall": 0.8483083749306711,
+      "eval_runtime": 44.876,
+      "eval_samples_per_second": 9.56,
+      "eval_steps_per_second": 1.604,
+      "step": 360
+    },
+    {
+      "epoch": 1.24,
+      "eval_accuracy": 0.9930487384256188,
+      "eval_f1": 0.8813077469793889,
+      "eval_loss": 0.03254028782248497,
+      "eval_precision": 0.9040536599591718,
+      "eval_recall": 0.8596783139212424,
+      "eval_runtime": 44.7683,
+      "eval_samples_per_second": 9.583,
+      "eval_steps_per_second": 1.608,
+      "step": 370
+    },
+    {
+      "epoch": 1.27,
+      "eval_accuracy": 0.9927038285001725,
+      "eval_f1": 0.8627060830017056,
+      "eval_loss": 0.030650299042463303,
+      "eval_precision": 0.8848396501457726,
+      "eval_recall": 0.8416528008874099,
+      "eval_runtime": 44.7059,
+      "eval_samples_per_second": 9.596,
+      "eval_steps_per_second": 1.611,
+      "step": 380
+    },
+    {
+      "epoch": 1.3,
+      "eval_accuracy": 0.9934467114165185,
+      "eval_f1": 0.8900583131844688,
+      "eval_loss": 0.029178058728575706,
+      "eval_precision": 0.9135766423357664,
+      "eval_recall": 0.867720465890183,
+      "eval_runtime": 44.67,
+      "eval_samples_per_second": 9.604,
+      "eval_steps_per_second": 1.612,
+      "step": 390
+    },
+    {
+      "epoch": 1.34,
+      "eval_accuracy": 0.993300787986522,
+      "eval_f1": 0.8887311950042579,
+      "eval_loss": 0.02753848023712635,
+      "eval_precision": 0.9101744186046512,
+      "eval_recall": 0.8682750970604548,
+      "eval_runtime": 44.9312,
+      "eval_samples_per_second": 9.548,
+      "eval_steps_per_second": 1.602,
+      "step": 400
+    },
+    {
+      "epoch": 1.37,
+      "eval_accuracy": 0.9936456979119683,
+      "eval_f1": 0.894469525959368,
+      "eval_loss": 0.027282511815428734,
+      "eval_precision": 0.9103963239517519,
+      "eval_recall": 0.8790904048807543,
+      "eval_runtime": 44.6914,
+      "eval_samples_per_second": 9.599,
+      "eval_steps_per_second": 1.611,
+      "step": 410
+    },
+    {
+      "epoch": 1.4,
+      "eval_accuracy": 0.9922925897429095,
+      "eval_f1": 0.8725379853685987,
+      "eval_loss": 0.02893451787531376,
+      "eval_precision": 0.8854940034266133,
+      "eval_recall": 0.8599556295063783,
+      "eval_runtime": 44.7493,
+      "eval_samples_per_second": 9.587,
+      "eval_steps_per_second": 1.609,
+      "step": 420
+    },
+    {
+      "epoch": 1.44,
+      "eval_accuracy": 0.9938712159401448,
+      "eval_f1": 0.8959954719117023,
+      "eval_loss": 0.024990031495690346,
+      "eval_precision": 0.9147645189251662,
+      "eval_recall": 0.8779811425402108,
+      "eval_runtime": 44.6669,
+      "eval_samples_per_second": 9.604,
+      "eval_steps_per_second": 1.612,
+      "step": 430
+    },
+    {
+      "epoch": 1.47,
+      "eval_accuracy": 0.9938181528746916,
+      "eval_f1": 0.8929327126534067,
+      "eval_loss": 0.026491543278098106,
+      "eval_precision": 0.9086993970714901,
+      "eval_recall": 0.8777038269550749,
+      "eval_runtime": 44.7596,
+      "eval_samples_per_second": 9.585,
+      "eval_steps_per_second": 1.609,
+      "step": 440
+    },
+    {
+      "epoch": 1.51,
+      "eval_accuracy": 0.9945610357910376,
+      "eval_f1": 0.9141895681435781,
+      "eval_loss": 0.0247773639857769,
+      "eval_precision": 0.9245604083947816,
+      "eval_recall": 0.9040488075429839,
+      "eval_runtime": 44.7459,
+      "eval_samples_per_second": 9.587,
+      "eval_steps_per_second": 1.609,
+      "step": 450
+    },
+    {
+      "epoch": 1.54,
+      "eval_accuracy": 0.9943753150619511,
+      "eval_f1": 0.9076814658210007,
+      "eval_loss": 0.0243705864995718,
+      "eval_precision": 0.9229005445686443,
+      "eval_recall": 0.8929561841375485,
+      "eval_runtime": 44.7374,
+      "eval_samples_per_second": 9.589,
+      "eval_steps_per_second": 1.609,
+      "step": 460
+    },
+    {
+      "epoch": 1.57,
+      "eval_accuracy": 0.994534504258311,
+      "eval_f1": 0.9149444678757205,
+      "eval_loss": 0.023830989375710487,
+      "eval_precision": 0.9278585685771314,
+      "eval_recall": 0.9023849140321686,
+      "eval_runtime": 44.7741,
+      "eval_samples_per_second": 9.581,
+      "eval_steps_per_second": 1.608,
+      "step": 470
+    },
+    {
+      "epoch": 1.61,
+      "eval_accuracy": 0.9947069592210342,
+      "eval_f1": 0.9177099451399635,
+      "eval_loss": 0.022406980395317078,
+      "eval_precision": 0.931201827005424,
+      "eval_recall": 0.9046034387132557,
+      "eval_runtime": 44.7025,
+      "eval_samples_per_second": 9.597,
+      "eval_steps_per_second": 1.611,
+      "step": 480
+    },
+    {
+      "epoch": 1.64,
+      "eval_accuracy": 0.9947069592210342,
+      "eval_f1": 0.9264129826524903,
+      "eval_loss": 0.023025579750537872,
+      "eval_precision": 0.9347826086956522,
+      "eval_recall": 0.918191902384914,
+      "eval_runtime": 45.0649,
+      "eval_samples_per_second": 9.52,
+      "eval_steps_per_second": 1.598,
+      "step": 490
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 9.5e-06,
+      "loss": 0.1784,
+      "step": 500
+    },
+    {
+      "epoch": 1.67,
+      "eval_accuracy": 0.9953702475392003,
+      "eval_f1": 0.9400780379041248,
+      "eval_loss": 0.022061465308070183,
+      "eval_precision": 0.9448179271708683,
+      "eval_recall": 0.9353854686633388,
+      "eval_runtime": 44.7675,
+      "eval_samples_per_second": 9.583,
+      "eval_steps_per_second": 1.608,
+      "step": 500
+    },
+    {
+      "epoch": 1.71,
+      "eval_accuracy": 0.9954100448382903,
+      "eval_f1": 0.9358616843279419,
+      "eval_loss": 0.02172040194272995,
+      "eval_precision": 0.9411104879416713,
+      "eval_recall": 0.9306711037160288,
+      "eval_runtime": 44.9669,
+      "eval_samples_per_second": 9.54,
+      "eval_steps_per_second": 1.601,
+      "step": 510
+    },
+    {
+      "epoch": 1.74,
+      "eval_accuracy": 0.9952243241092038,
+      "eval_f1": 0.9297780259667735,
+      "eval_loss": 0.021418150514364243,
+      "eval_precision": 0.9361821759910036,
+      "eval_recall": 0.9234608985024958,
+      "eval_runtime": 44.7829,
+      "eval_samples_per_second": 9.58,
+      "eval_steps_per_second": 1.608,
+      "step": 520
+    },
+    {
+      "epoch": 1.77,
+      "eval_accuracy": 0.9950651349128439,
+      "eval_f1": 0.9294002516426674,
+      "eval_loss": 0.02137417532503605,
+      "eval_precision": 0.937129968987877,
+      "eval_recall": 0.9217970049916805,
+      "eval_runtime": 44.837,
+      "eval_samples_per_second": 9.568,
+      "eval_steps_per_second": 1.606,
+      "step": 530
+    },
+    {
+      "epoch": 1.81,
+      "eval_accuracy": 0.9954100448382903,
+      "eval_f1": 0.9379098646574578,
+      "eval_loss": 0.02044159732758999,
+      "eval_precision": 0.94383600112328,
+      "eval_recall": 0.9320576816417082,
+      "eval_runtime": 44.8548,
+      "eval_samples_per_second": 9.564,
+      "eval_steps_per_second": 1.605,
+      "step": 540
+    },
+    {
+      "epoch": 1.84,
+      "eval_accuracy": 0.9959274097264599,
+      "eval_f1": 0.9561111111111111,
+      "eval_loss": 0.020166810601949692,
+      "eval_precision": 0.9577072899276572,
+      "eval_recall": 0.954520244037715,
+      "eval_runtime": 44.8166,
+      "eval_samples_per_second": 9.572,
+      "eval_steps_per_second": 1.607,
+      "step": 550
+    },
+    {
+      "epoch": 1.87,
+      "eval_accuracy": 0.9952375898755671,
+      "eval_f1": 0.9289403881055422,
+      "eval_loss": 0.020729683339595795,
+      "eval_precision": 0.935338768625246,
+      "eval_recall": 0.9226289517470881,
+      "eval_runtime": 44.8369,
+      "eval_samples_per_second": 9.568,
+      "eval_steps_per_second": 1.606,
+      "step": 560
+    },
+    {
+      "epoch": 1.91,
+      "eval_accuracy": 0.9960600673900931,
+      "eval_f1": 0.9606211869107044,
+      "eval_loss": 0.019887683913111687,
+      "eval_precision": 0.9606211869107044,
+      "eval_recall": 0.9606211869107044,
+      "eval_runtime": 44.8366,
+      "eval_samples_per_second": 9.568,
+      "eval_steps_per_second": 1.606,
+      "step": 570
+    },
+    {
+      "epoch": 1.94,
+      "eval_accuracy": 0.9956753601655568,
+      "eval_f1": 0.9422407794015311,
+      "eval_loss": 0.021609965711832047,
+      "eval_precision": 0.945794914780665,
+      "eval_recall": 0.9387132556849695,
+      "eval_runtime": 44.9147,
+      "eval_samples_per_second": 9.551,
+      "eval_steps_per_second": 1.603,
+      "step": 580
+    },
+    {
+      "epoch": 1.97,
+      "eval_accuracy": 0.9960468016237298,
+      "eval_f1": 0.9564975677553856,
+      "eval_loss": 0.019716205075383186,
+      "eval_precision": 0.9587628865979382,
+      "eval_recall": 0.9542429284525791,
+      "eval_runtime": 45.3593,
+      "eval_samples_per_second": 9.458,
+      "eval_steps_per_second": 1.587,
+      "step": 590
+    },
+    {
+      "epoch": 2.01,
+      "eval_accuracy": 0.9957549547637367,
+      "eval_f1": 0.9439577249339453,
+      "eval_loss": 0.020143885165452957,
+      "eval_precision": 0.9467224546722455,
+      "eval_recall": 0.9412090959511925,
+      "eval_runtime": 45.0377,
+      "eval_samples_per_second": 9.525,
+      "eval_steps_per_second": 1.599,
+      "step": 600
+    },
+    {
+      "epoch": 2.04,
+      "eval_accuracy": 0.9955029052028336,
+      "eval_f1": 0.9359207036158034,
+      "eval_loss": 0.019947798922657967,
+      "eval_precision": 0.9423671633398931,
+      "eval_recall": 0.9295618413754853,
+      "eval_runtime": 45.1113,
+      "eval_samples_per_second": 9.51,
+      "eval_steps_per_second": 1.596,
+      "step": 610
+    },
+    {
+      "epoch": 2.07,
+      "eval_accuracy": 0.9961131304555464,
+      "eval_f1": 0.9537564227190668,
+      "eval_loss": 0.01930239051580429,
+      "eval_precision": 0.9552155771905424,
+      "eval_recall": 0.9523017193566279,
+      "eval_runtime": 45.0399,
+      "eval_samples_per_second": 9.525,
+      "eval_steps_per_second": 1.599,
+      "step": 620
+    },
+    {
+      "epoch": 2.11,
+      "eval_accuracy": 0.9959141439600966,
+      "eval_f1": 0.9510664993726474,
+      "eval_loss": 0.020125582814216614,
+      "eval_precision": 0.9562657695542472,
+      "eval_recall": 0.9459234608985025,
+      "eval_runtime": 44.9443,
+      "eval_samples_per_second": 9.545,
+      "eval_steps_per_second": 1.602,
+      "step": 630
+    },
+    {
+      "epoch": 2.14,
+      "eval_accuracy": 0.9965376349791727,
+      "eval_f1": 0.9709141274238227,
+      "eval_loss": 0.01933199167251587,
+      "eval_precision": 0.9698395130049806,
+      "eval_recall": 0.9719911259012757,
+      "eval_runtime": 44.9595,
+      "eval_samples_per_second": 9.542,
+      "eval_steps_per_second": 1.601,
+      "step": 640
+    },
+    {
+      "epoch": 2.17,
+      "eval_accuracy": 0.9962988511846329,
+      "eval_f1": 0.9642262895174709,
+      "eval_loss": 0.0191953107714653,
+      "eval_precision": 0.9642262895174709,
+      "eval_recall": 0.9642262895174709,
+      "eval_runtime": 44.8109,
+      "eval_samples_per_second": 9.574,
+      "eval_steps_per_second": 1.607,
+      "step": 650
+    },
+    {
+      "epoch": 2.21,
+      "eval_accuracy": 0.9957814862964633,
+      "eval_f1": 0.947662247034194,
+      "eval_loss": 0.019274834543466568,
+      "eval_precision": 0.9539196403484125,
+      "eval_recall": 0.9414864115363284,
+      "eval_runtime": 45.2879,
+      "eval_samples_per_second": 9.473,
+      "eval_steps_per_second": 1.59,
+      "step": 660
+    },
+    {
+      "epoch": 2.24,
+      "eval_accuracy": 0.9965906980446261,
+      "eval_f1": 0.9721568084222192,
+      "eval_loss": 0.018348801881074905,
+      "eval_precision": 0.9712150567395517,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 44.979,
+      "eval_samples_per_second": 9.538,
+      "eval_steps_per_second": 1.601,
+      "step": 670
+    },
+    {
+      "epoch": 2.27,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.975387168141593,
+      "eval_loss": 0.018433524295687675,
+      "eval_precision": 0.9726971869829013,
+      "eval_recall": 0.9780920687742651,
+      "eval_runtime": 44.9845,
+      "eval_samples_per_second": 9.537,
+      "eval_steps_per_second": 1.601,
+      "step": 680
+    },
+    {
+      "epoch": 2.31,
+      "eval_accuracy": 0.9967631530073492,
+      "eval_f1": 0.9695880564003317,
+      "eval_loss": 0.018039193004369736,
+      "eval_precision": 0.9666482910694597,
+      "eval_recall": 0.9725457570715474,
+      "eval_runtime": 44.9711,
+      "eval_samples_per_second": 9.539,
+      "eval_steps_per_second": 1.601,
+      "step": 690
+    },
+    {
+      "epoch": 2.34,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9738625363020329,
+      "eval_loss": 0.01683180034160614,
+      "eval_precision": 0.9713103448275862,
+      "eval_recall": 0.9764281752634498,
+      "eval_runtime": 44.9102,
+      "eval_samples_per_second": 9.552,
+      "eval_steps_per_second": 1.603,
+      "step": 700
+    },
+    {
+      "epoch": 2.37,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9755152856550007,
+      "eval_loss": 0.01656004600226879,
+      "eval_precision": 0.9732266077836047,
+      "eval_recall": 0.9778147531891292,
+      "eval_runtime": 44.9636,
+      "eval_samples_per_second": 9.541,
+      "eval_steps_per_second": 1.601,
+      "step": 710
+    },
+    {
+      "epoch": 2.41,
+      "eval_accuracy": 0.9967498872409859,
+      "eval_f1": 0.973655019412091,
+      "eval_loss": 0.016868896782398224,
+      "eval_precision": 0.973655019412091,
+      "eval_recall": 0.973655019412091,
+      "eval_runtime": 44.8793,
+      "eval_samples_per_second": 9.559,
+      "eval_steps_per_second": 1.604,
+      "step": 720
+    },
+    {
+      "epoch": 2.44,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.975765129483451,
+      "eval_loss": 0.01695641689002514,
+      "eval_precision": 0.9745504840940525,
+      "eval_recall": 0.9769828064337216,
+      "eval_runtime": 44.9769,
+      "eval_samples_per_second": 9.538,
+      "eval_steps_per_second": 1.601,
+      "step": 730
+    },
+    {
+      "epoch": 2.47,
+      "eval_accuracy": 0.9966968241755326,
+      "eval_f1": 0.9696128763701957,
+      "eval_loss": 0.017096424475312233,
+      "eval_precision": 0.9702860316578729,
+      "eval_recall": 0.968940654464781,
+      "eval_runtime": 45.2584,
+      "eval_samples_per_second": 9.479,
+      "eval_steps_per_second": 1.591,
+      "step": 740
+    },
+    {
+      "epoch": 2.51,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.9747992245915259,
+      "eval_loss": 0.016900014132261276,
+      "eval_precision": 0.9734513274336283,
+      "eval_recall": 0.9761508596783139,
+      "eval_runtime": 44.9939,
+      "eval_samples_per_second": 9.535,
+      "eval_steps_per_second": 1.6,
+      "step": 750
+    },
+    {
+      "epoch": 2.54,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9747992245915259,
+      "eval_loss": 0.01711142621934414,
+      "eval_precision": 0.9734513274336283,
+      "eval_recall": 0.9761508596783139,
+      "eval_runtime": 44.9916,
+      "eval_samples_per_second": 9.535,
+      "eval_steps_per_second": 1.6,
+      "step": 760
+    },
+    {
+      "epoch": 2.58,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9726957726957727,
+      "eval_loss": 0.017322950065135956,
+      "eval_precision": 0.972291493488501,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 44.9667,
+      "eval_samples_per_second": 9.54,
+      "eval_steps_per_second": 1.601,
+      "step": 770
+    },
+    {
+      "epoch": 2.61,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9776853776853778,
+      "eval_loss": 0.015976430848240852,
+      "eval_precision": 0.9772790246605708,
+      "eval_recall": 0.9780920687742651,
+      "eval_runtime": 44.9823,
+      "eval_samples_per_second": 9.537,
+      "eval_steps_per_second": 1.601,
+      "step": 780
+    },
+    {
+      "epoch": 2.64,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.9725685785536159,
+      "eval_loss": 0.016221042722463608,
+      "eval_precision": 0.9717607973421927,
+      "eval_recall": 0.9733777038269551,
+      "eval_runtime": 45.0356,
+      "eval_samples_per_second": 9.526,
+      "eval_steps_per_second": 1.599,
+      "step": 790
+    },
+    {
+      "epoch": 2.68,
+      "eval_accuracy": 0.9967764187737126,
+      "eval_f1": 0.9714760454167821,
+      "eval_loss": 0.016350209712982178,
+      "eval_precision": 0.9701327433628318,
+      "eval_recall": 0.9728230726566833,
+      "eval_runtime": 44.9594,
+      "eval_samples_per_second": 9.542,
+      "eval_steps_per_second": 1.601,
+      "step": 800
+    },
+    {
+      "epoch": 2.71,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9753325942350333,
+      "eval_loss": 0.0164735559374094,
+      "eval_precision": 0.974792243767313,
+      "eval_recall": 0.9758735440931781,
+      "eval_runtime": 44.8809,
+      "eval_samples_per_second": 9.559,
+      "eval_steps_per_second": 1.604,
+      "step": 810
+    },
+    {
+      "epoch": 2.74,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9754541672444876,
+      "eval_loss": 0.01638418808579445,
+      "eval_precision": 0.9755894590846047,
+      "eval_recall": 0.9753189129229063,
+      "eval_runtime": 45.2809,
+      "eval_samples_per_second": 9.474,
+      "eval_steps_per_second": 1.59,
+      "step": 820
+    },
+    {
+      "epoch": 2.78,
+      "eval_accuracy": 0.9951712610437505,
+      "eval_f1": 0.9239864864864865,
+      "eval_loss": 0.0193793848156929,
+      "eval_precision": 0.9382504288164666,
+      "eval_recall": 0.9101497504159733,
+      "eval_runtime": 44.9984,
+      "eval_samples_per_second": 9.534,
+      "eval_steps_per_second": 1.6,
+      "step": 830
+    },
+    {
+      "epoch": 2.81,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.979071379071379,
+      "eval_loss": 0.015317755751311779,
+      "eval_precision": 0.9786644499861458,
+      "eval_recall": 0.9794786466999446,
+      "eval_runtime": 44.8121,
+      "eval_samples_per_second": 9.573,
+      "eval_steps_per_second": 1.607,
+      "step": 840
+    },
+    {
+      "epoch": 2.84,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9747712780704185,
+      "eval_loss": 0.016130488365888596,
+      "eval_precision": 0.9745011086474501,
+      "eval_recall": 0.9750415973377704,
+      "eval_runtime": 44.9246,
+      "eval_samples_per_second": 9.549,
+      "eval_steps_per_second": 1.603,
+      "step": 850
+    },
+    {
+      "epoch": 2.88,
+      "eval_accuracy": 0.9965111034464461,
+      "eval_f1": 0.9585304759254106,
+      "eval_loss": 0.017094574868679047,
+      "eval_precision": 0.9620111731843576,
+      "eval_recall": 0.9550748752079867,
+      "eval_runtime": 44.9689,
+      "eval_samples_per_second": 9.54,
+      "eval_steps_per_second": 1.601,
+      "step": 860
+    },
+    {
+      "epoch": 2.91,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9767634854771784,
+      "eval_loss": 0.016324549913406372,
+      "eval_precision": 0.9743377483443708,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 44.9212,
+      "eval_samples_per_second": 9.55,
+      "eval_steps_per_second": 1.603,
+      "step": 870
+    },
+    {
+      "epoch": 2.94,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9753394292047659,
+      "eval_loss": 0.016577888280153275,
+      "eval_precision": 0.9745293466223699,
+      "eval_recall": 0.9761508596783139,
+      "eval_runtime": 45.0626,
+      "eval_samples_per_second": 9.52,
+      "eval_steps_per_second": 1.598,
+      "step": 880
+    },
+    {
+      "epoch": 2.98,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9743589743589743,
+      "eval_loss": 0.01666262373328209,
+      "eval_precision": 0.973954003879191,
+      "eval_recall": 0.9747642817526345,
+      "eval_runtime": 44.929,
+      "eval_samples_per_second": 9.548,
+      "eval_steps_per_second": 1.603,
+      "step": 890
+    },
+    {
+      "epoch": 3.01,
+      "eval_accuracy": 0.9966570268764426,
+      "eval_f1": 0.9715396362626684,
+      "eval_loss": 0.01778605580329895,
+      "eval_precision": 0.9727550736725049,
+      "eval_recall": 0.9703272323904604,
+      "eval_runtime": 44.9166,
+      "eval_samples_per_second": 9.551,
+      "eval_steps_per_second": 1.603,
+      "step": 900
+    },
+    {
+      "epoch": 3.04,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9786407766990292,
+      "eval_loss": 0.016635755077004433,
+      "eval_precision": 0.978912319644839,
+      "eval_recall": 0.978369384359401,
+      "eval_runtime": 45.1716,
+      "eval_samples_per_second": 9.497,
+      "eval_steps_per_second": 1.594,
+      "step": 910
+    },
+    {
+      "epoch": 3.08,
+      "eval_accuracy": 0.9971478602318856,
+      "eval_f1": 0.9811529933481152,
+      "eval_loss": 0.01662967912852764,
+      "eval_precision": 0.9806094182825484,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 44.9112,
+      "eval_samples_per_second": 9.552,
+      "eval_steps_per_second": 1.603,
+      "step": 920
+    },
+    {
+      "epoch": 3.11,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9753052164261932,
+      "eval_loss": 0.016620052978396416,
+      "eval_precision": 0.975846751804553,
+      "eval_recall": 0.9747642817526345,
+      "eval_runtime": 44.9363,
+      "eval_samples_per_second": 9.547,
+      "eval_steps_per_second": 1.602,
+      "step": 930
+    },
+    {
+      "epoch": 3.14,
+      "eval_accuracy": 0.9972407205964289,
+      "eval_f1": 0.9815508392287419,
+      "eval_loss": 0.015619627200067043,
+      "eval_precision": 0.9819594782126007,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 44.9174,
+      "eval_samples_per_second": 9.551,
+      "eval_steps_per_second": 1.603,
+      "step": 940
+    },
+    {
+      "epoch": 3.18,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9779442363712026,
+      "eval_loss": 0.015341303311288357,
+      "eval_precision": 0.9783513738551207,
+      "eval_recall": 0.9775374376039934,
+      "eval_runtime": 45.0123,
+      "eval_samples_per_second": 9.531,
+      "eval_steps_per_second": 1.6,
+      "step": 950
+    },
+    {
+      "epoch": 3.21,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9764281752634498,
+      "eval_loss": 0.01650950312614441,
+      "eval_precision": 0.9764281752634498,
+      "eval_recall": 0.9764281752634498,
+      "eval_runtime": 44.93,
+      "eval_samples_per_second": 9.548,
+      "eval_steps_per_second": 1.602,
+      "step": 960
+    },
+    {
+      "epoch": 3.24,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9774144381321879,
+      "eval_loss": 0.01659621112048626,
+      "eval_precision": 0.9767377457767932,
+      "eval_recall": 0.9780920687742651,
+      "eval_runtime": 44.9897,
+      "eval_samples_per_second": 9.536,
+      "eval_steps_per_second": 1.6,
+      "step": 970
+    },
+    {
+      "epoch": 3.28,
+      "eval_accuracy": 0.9967498872409859,
+      "eval_f1": 0.9746572496884087,
+      "eval_loss": 0.016826625913381577,
+      "eval_precision": 0.9734439834024896,
+      "eval_recall": 0.9758735440931781,
+      "eval_runtime": 45.2137,
+      "eval_samples_per_second": 9.488,
+      "eval_steps_per_second": 1.592,
+      "step": 980
+    },
+    {
+      "epoch": 3.31,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.978146611341632,
+      "eval_loss": 0.016263196244835854,
+      "eval_precision": 0.9757174392935982,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 44.9271,
+      "eval_samples_per_second": 9.549,
+      "eval_steps_per_second": 1.603,
+      "step": 990
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9e-06,
+      "loss": 0.0154,
+      "step": 1000
+    },
+    {
+      "epoch": 3.34,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.9775809576529201,
+      "eval_loss": 0.016901057213544846,
+      "eval_precision": 0.9756906077348066,
+      "eval_recall": 0.9794786466999446,
+      "eval_runtime": 44.996,
+      "eval_samples_per_second": 9.534,
+      "eval_steps_per_second": 1.6,
+      "step": 1000
+    },
+    {
+      "epoch": 3.38,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9803269603768356,
+      "eval_loss": 0.01609906367957592,
+      "eval_precision": 0.9795127353266888,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.3202,
+      "eval_samples_per_second": 9.466,
+      "eval_steps_per_second": 1.589,
+      "step": 1010
+    },
+    {
+      "epoch": 3.41,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.977993079584775,
+      "eval_loss": 0.016519052907824516,
+      "eval_precision": 0.9762365294280188,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.0574,
+      "eval_samples_per_second": 9.521,
+      "eval_steps_per_second": 1.598,
+      "step": 1020
+    },
+    {
+      "epoch": 3.44,
+      "eval_accuracy": 0.9967233557082593,
+      "eval_f1": 0.9722530521642619,
+      "eval_loss": 0.01666153408586979,
+      "eval_precision": 0.9727928928373126,
+      "eval_recall": 0.9717138103161398,
+      "eval_runtime": 45.0201,
+      "eval_samples_per_second": 9.529,
+      "eval_steps_per_second": 1.599,
+      "step": 1030
+    },
+    {
+      "epoch": 3.48,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9761640798226164,
+      "eval_loss": 0.015780288726091385,
+      "eval_precision": 0.9756232686980609,
+      "eval_recall": 0.9767054908485857,
+      "eval_runtime": 45.2347,
+      "eval_samples_per_second": 9.484,
+      "eval_steps_per_second": 1.592,
+      "step": 1040
+    },
+    {
+      "epoch": 3.51,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9766284054764209,
+      "eval_loss": 0.015381171368062496,
+      "eval_precision": 0.9740689655172414,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 45.2769,
+      "eval_samples_per_second": 9.475,
+      "eval_steps_per_second": 1.59,
+      "step": 1050
+    },
+    {
+      "epoch": 3.55,
+      "eval_accuracy": 0.9959804727919132,
+      "eval_f1": 0.9448555074689375,
+      "eval_loss": 0.01652899943292141,
+      "eval_precision": 0.9513635085746416,
+      "eval_recall": 0.9384359400998337,
+      "eval_runtime": 45.1375,
+      "eval_samples_per_second": 9.504,
+      "eval_steps_per_second": 1.595,
+      "step": 1060
+    },
+    {
+      "epoch": 3.58,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9749619587771476,
+      "eval_loss": 0.015719007700681686,
+      "eval_precision": 0.9726745790781121,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.0215,
+      "eval_samples_per_second": 9.529,
+      "eval_steps_per_second": 1.599,
+      "step": 1070
+    },
+    {
+      "epoch": 3.61,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9782458085076902,
+      "eval_loss": 0.01600920408964157,
+      "eval_precision": 0.977568540570479,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.1658,
+      "eval_samples_per_second": 9.498,
+      "eval_steps_per_second": 1.594,
+      "step": 1080
+    },
+    {
+      "epoch": 3.65,
+      "eval_accuracy": 0.99580801782919,
+      "eval_f1": 0.9429732380552053,
+      "eval_loss": 0.01786983013153076,
+      "eval_precision": 0.9529878221467006,
+      "eval_recall": 0.9331669439822518,
+      "eval_runtime": 45.14,
+      "eval_samples_per_second": 9.504,
+      "eval_steps_per_second": 1.595,
+      "step": 1090
+    },
+    {
+      "epoch": 3.68,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.977439446366782,
+      "eval_loss": 0.015765171498060226,
+      "eval_precision": 0.9756838905775076,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 45.1016,
+      "eval_samples_per_second": 9.512,
+      "eval_steps_per_second": 1.596,
+      "step": 1100
+    },
+    {
+      "epoch": 3.71,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.977703919124775,
+      "eval_loss": 0.015863990411162376,
+      "eval_precision": 0.9764868603042877,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.1212,
+      "eval_samples_per_second": 9.508,
+      "eval_steps_per_second": 1.596,
+      "step": 1110
+    },
+    {
+      "epoch": 3.75,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.975609756097561,
+      "eval_loss": 0.01584913209080696,
+      "eval_precision": 0.9750692520775623,
+      "eval_recall": 0.9761508596783139,
+      "eval_runtime": 45.0612,
+      "eval_samples_per_second": 9.52,
+      "eval_steps_per_second": 1.598,
+      "step": 1120
+    },
+    {
+      "epoch": 3.78,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9744869661674986,
+      "eval_loss": 0.01588474027812481,
+      "eval_precision": 0.9744869661674986,
+      "eval_recall": 0.9744869661674986,
+      "eval_runtime": 45.3578,
+      "eval_samples_per_second": 9.458,
+      "eval_steps_per_second": 1.587,
+      "step": 1130
+    },
+    {
+      "epoch": 3.81,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9752846431546792,
+      "eval_loss": 0.016261184588074684,
+      "eval_precision": 0.9766407119021134,
+      "eval_recall": 0.9739323349972269,
+      "eval_runtime": 45.1022,
+      "eval_samples_per_second": 9.512,
+      "eval_steps_per_second": 1.596,
+      "step": 1140
+    },
+    {
+      "epoch": 3.85,
+      "eval_accuracy": 0.9967100899418959,
+      "eval_f1": 0.9723145071982281,
+      "eval_loss": 0.016805831342935562,
+      "eval_precision": 0.970702045328911,
+      "eval_recall": 0.9739323349972269,
+      "eval_runtime": 45.1109,
+      "eval_samples_per_second": 9.51,
+      "eval_steps_per_second": 1.596,
+      "step": 1150
+    },
+    {
+      "epoch": 3.88,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9805771365149833,
+      "eval_loss": 0.01488415990024805,
+      "eval_precision": 0.9811215991116047,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.1408,
+      "eval_samples_per_second": 9.504,
+      "eval_steps_per_second": 1.595,
+      "step": 1160
+    },
+    {
+      "epoch": 3.91,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9792531120331951,
+      "eval_loss": 0.014714999124407768,
+      "eval_precision": 0.9768211920529801,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.0991,
+      "eval_samples_per_second": 9.512,
+      "eval_steps_per_second": 1.596,
+      "step": 1170
+    },
+    {
+      "epoch": 3.95,
+      "eval_accuracy": 0.9969754052691624,
+      "eval_f1": 0.9757986447241046,
+      "eval_loss": 0.01566295139491558,
+      "eval_precision": 0.9732413793103448,
+      "eval_recall": 0.978369384359401,
+      "eval_runtime": 45.1028,
+      "eval_samples_per_second": 9.512,
+      "eval_steps_per_second": 1.596,
+      "step": 1180
+    },
+    {
+      "epoch": 3.98,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9786880708552449,
+      "eval_loss": 0.015323741361498833,
+      "eval_precision": 0.9767955801104973,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.1524,
+      "eval_samples_per_second": 9.501,
+      "eval_steps_per_second": 1.595,
+      "step": 1190
+    },
+    {
+      "epoch": 4.01,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9747642817526345,
+      "eval_loss": 0.015690365806221962,
+      "eval_precision": 0.9747642817526345,
+      "eval_recall": 0.9747642817526345,
+      "eval_runtime": 45.1208,
+      "eval_samples_per_second": 9.508,
+      "eval_steps_per_second": 1.596,
+      "step": 1200
+    },
+    {
+      "epoch": 4.05,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9797783933518006,
+      "eval_loss": 0.015430403873324394,
+      "eval_precision": 0.978693967902601,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.032,
+      "eval_samples_per_second": 9.527,
+      "eval_steps_per_second": 1.599,
+      "step": 1210
+    },
+    {
+      "epoch": 4.08,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9780052566053397,
+      "eval_loss": 0.0156533382833004,
+      "eval_precision": 0.9757107369583218,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.2988,
+      "eval_samples_per_second": 9.47,
+      "eval_steps_per_second": 1.589,
+      "step": 1220
+    },
+    {
+      "epoch": 4.11,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9735493698933665,
+      "eval_loss": 0.016326196491718292,
+      "eval_precision": 0.9723374827109267,
+      "eval_recall": 0.9747642817526345,
+      "eval_runtime": 45.0756,
+      "eval_samples_per_second": 9.517,
+      "eval_steps_per_second": 1.597,
+      "step": 1230
+    },
+    {
+      "epoch": 4.15,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9778699861687413,
+      "eval_loss": 0.015600275248289108,
+      "eval_precision": 0.9754415011037527,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 44.9579,
+      "eval_samples_per_second": 9.542,
+      "eval_steps_per_second": 1.601,
+      "step": 1240
+    },
+    {
+      "epoch": 4.18,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9774331995015921,
+      "eval_loss": 0.015862880274653435,
+      "eval_precision": 0.9759469173348079,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.0784,
+      "eval_samples_per_second": 9.517,
+      "eval_steps_per_second": 1.597,
+      "step": 1250
+    },
+    {
+      "epoch": 4.21,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9802076124567475,
+      "eval_loss": 0.01617247611284256,
+      "eval_precision": 0.9784470848300636,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.0516,
+      "eval_samples_per_second": 9.522,
+      "eval_steps_per_second": 1.598,
+      "step": 1260
+    },
+    {
+      "epoch": 4.25,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.978146611341632,
+      "eval_loss": 0.016013789921998978,
+      "eval_precision": 0.9757174392935982,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.1173,
+      "eval_samples_per_second": 9.509,
+      "eval_steps_per_second": 1.596,
+      "step": 1270
+    },
+    {
+      "epoch": 4.28,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9744157101369105,
+      "eval_loss": 0.01655399613082409,
+      "eval_precision": 0.9718620689655172,
+      "eval_recall": 0.9769828064337216,
+      "eval_runtime": 45.0171,
+      "eval_samples_per_second": 9.53,
+      "eval_steps_per_second": 1.599,
+      "step": 1280
+    },
+    {
+      "epoch": 4.31,
+      "eval_accuracy": 0.9971080629327956,
+      "eval_f1": 0.9793600221637346,
+      "eval_loss": 0.015200940892100334,
+      "eval_precision": 0.9784112925546637,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.0877,
+      "eval_samples_per_second": 9.515,
+      "eval_steps_per_second": 1.597,
+      "step": 1290
+    },
+    {
+      "epoch": 4.35,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.9751175006911805,
+      "eval_loss": 0.016674669459462166,
+      "eval_precision": 0.9721609702315325,
+      "eval_recall": 0.9780920687742651,
+      "eval_runtime": 45.3184,
+      "eval_samples_per_second": 9.466,
+      "eval_steps_per_second": 1.589,
+      "step": 1300
+    },
+    {
+      "epoch": 4.38,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.9766413268832067,
+      "eval_loss": 0.01657554879784584,
+      "eval_precision": 0.9735464315238358,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.1144,
+      "eval_samples_per_second": 9.509,
+      "eval_steps_per_second": 1.596,
+      "step": 1310
+    },
+    {
+      "epoch": 4.41,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9781586950511474,
+      "eval_loss": 0.016078708693385124,
+      "eval_precision": 0.9751929437706726,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.1076,
+      "eval_samples_per_second": 9.511,
+      "eval_steps_per_second": 1.596,
+      "step": 1320
+    },
+    {
+      "epoch": 4.45,
+      "eval_accuracy": 0.9967764187737126,
+      "eval_f1": 0.9739251040221916,
+      "eval_loss": 0.015717538073658943,
+      "eval_precision": 0.9741953385127636,
+      "eval_recall": 0.973655019412091,
+      "eval_runtime": 45.0205,
+      "eval_samples_per_second": 9.529,
+      "eval_steps_per_second": 1.599,
+      "step": 1330
+    },
+    {
+      "epoch": 4.48,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9795127353266887,
+      "eval_loss": 0.015196431428194046,
+      "eval_precision": 0.9778883360972913,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.4378,
+      "eval_samples_per_second": 9.441,
+      "eval_steps_per_second": 1.585,
+      "step": 1340
+    },
+    {
+      "epoch": 4.52,
+      "eval_accuracy": 0.9967233557082593,
+      "eval_f1": 0.971064654575661,
+      "eval_loss": 0.015703538432717323,
+      "eval_precision": 0.9695880564003317,
+      "eval_recall": 0.9725457570715474,
+      "eval_runtime": 45.4022,
+      "eval_samples_per_second": 9.449,
+      "eval_steps_per_second": 1.586,
+      "step": 1350
+    },
+    {
+      "epoch": 4.55,
+      "eval_accuracy": 0.9968162160728026,
+      "eval_f1": 0.9755828390122776,
+      "eval_loss": 0.015339327044785023,
+      "eval_precision": 0.9706286027998902,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.4342,
+      "eval_samples_per_second": 9.442,
+      "eval_steps_per_second": 1.585,
+      "step": 1360
+    },
+    {
+      "epoch": 4.58,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9758587391364325,
+      "eval_loss": 0.015390865504741669,
+      "eval_precision": 0.9709031018391435,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.7337,
+      "eval_samples_per_second": 9.38,
+      "eval_steps_per_second": 1.574,
+      "step": 1370
+    },
+    {
+      "epoch": 4.62,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9785822854774078,
+      "eval_loss": 0.015035979449748993,
+      "eval_precision": 0.97521343982374,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.4663,
+      "eval_samples_per_second": 9.436,
+      "eval_steps_per_second": 1.584,
+      "step": 1380
+    },
+    {
+      "epoch": 4.65,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9809234171965718,
+      "eval_loss": 0.014896390959620476,
+      "eval_precision": 0.9779492833517089,
+      "eval_recall": 0.9839156960621187,
+      "eval_runtime": 45.0989,
+      "eval_samples_per_second": 9.512,
+      "eval_steps_per_second": 1.596,
+      "step": 1390
+    },
+    {
+      "epoch": 4.68,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9768339768339769,
+      "eval_loss": 0.015806537121534348,
+      "eval_precision": 0.9714755896873286,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.0416,
+      "eval_samples_per_second": 9.525,
+      "eval_steps_per_second": 1.599,
+      "step": 1400
+    },
+    {
+      "epoch": 4.72,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9778761061946902,
+      "eval_loss": 0.015140415169298649,
+      "eval_precision": 0.9751792608935466,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.6611,
+      "eval_samples_per_second": 9.395,
+      "eval_steps_per_second": 1.577,
+      "step": 1410
+    },
+    {
+      "epoch": 4.75,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9795353982300885,
+      "eval_loss": 0.015615841373801231,
+      "eval_precision": 0.9768339768339769,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.3516,
+      "eval_samples_per_second": 9.459,
+      "eval_steps_per_second": 1.588,
+      "step": 1420
+    },
+    {
+      "epoch": 4.78,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.9802240354031254,
+      "eval_loss": 0.015609286725521088,
+      "eval_precision": 0.9776551724137931,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.8817,
+      "eval_samples_per_second": 9.35,
+      "eval_steps_per_second": 1.569,
+      "step": 1430
+    },
+    {
+      "epoch": 4.82,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.980777209238003,
+      "eval_loss": 0.015975775197148323,
+      "eval_precision": 0.9782068965517241,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.1602,
+      "eval_samples_per_second": 9.5,
+      "eval_steps_per_second": 1.594,
+      "step": 1440
+    },
+    {
+      "epoch": 4.85,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9810642709053213,
+      "eval_loss": 0.015114562585949898,
+      "eval_precision": 0.9779553596031965,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.0898,
+      "eval_samples_per_second": 9.514,
+      "eval_steps_per_second": 1.597,
+      "step": 1450
+    },
+    {
+      "epoch": 4.88,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9806576402321082,
+      "eval_loss": 0.01497586164623499,
+      "eval_precision": 0.977147577092511,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 47.111,
+      "eval_samples_per_second": 9.106,
+      "eval_steps_per_second": 1.528,
+      "step": 1460
+    },
+    {
+      "epoch": 4.92,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9779005524861878,
+      "eval_loss": 0.014804758131504059,
+      "eval_precision": 0.9741331865712713,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.7757,
+      "eval_samples_per_second": 9.372,
+      "eval_steps_per_second": 1.573,
+      "step": 1470
+    },
+    {
+      "epoch": 4.95,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9767634854771784,
+      "eval_loss": 0.015325279906392097,
+      "eval_precision": 0.9743377483443708,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 45.788,
+      "eval_samples_per_second": 9.369,
+      "eval_steps_per_second": 1.572,
+      "step": 1480
+    },
+    {
+      "epoch": 4.98,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9820143884892086,
+      "eval_loss": 0.014990455470979214,
+      "eval_precision": 0.9798453892876864,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.1213,
+      "eval_samples_per_second": 9.508,
+      "eval_steps_per_second": 1.596,
+      "step": 1490
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 8.5e-06,
+      "loss": 0.0081,
+      "step": 1500
+    },
+    {
+      "epoch": 5.02,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9796821008984106,
+      "eval_loss": 0.01517151016741991,
+      "eval_precision": 0.9765775695783963,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.193,
+      "eval_samples_per_second": 9.493,
+      "eval_steps_per_second": 1.593,
+      "step": 1500
+    },
+    {
+      "epoch": 5.05,
+      "eval_accuracy": 0.9967366214746226,
+      "eval_f1": 0.9682408174537421,
+      "eval_loss": 0.01580970361828804,
+      "eval_precision": 0.9642464246424642,
+      "eval_recall": 0.9722684414864116,
+      "eval_runtime": 45.4381,
+      "eval_samples_per_second": 9.441,
+      "eval_steps_per_second": 1.585,
+      "step": 1510
+    },
+    {
+      "epoch": 5.08,
+      "eval_accuracy": 0.9966702926428059,
+      "eval_f1": 0.9666759233546237,
+      "eval_loss": 0.01570785790681839,
+      "eval_precision": 0.9680200222469411,
+      "eval_recall": 0.9653355518580145,
+      "eval_runtime": 45.4491,
+      "eval_samples_per_second": 9.439,
+      "eval_steps_per_second": 1.584,
+      "step": 1520
+    },
+    {
+      "epoch": 5.12,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9800388134183532,
+      "eval_loss": 0.015114962123334408,
+      "eval_precision": 0.9797671840354767,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.1961,
+      "eval_samples_per_second": 9.492,
+      "eval_steps_per_second": 1.593,
+      "step": 1530
+    },
+    {
+      "epoch": 5.15,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9818835569077582,
+      "eval_loss": 0.01481586042791605,
+      "eval_precision": 0.9793103448275862,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.6531,
+      "eval_samples_per_second": 9.397,
+      "eval_steps_per_second": 1.577,
+      "step": 1540
+    },
+    {
+      "epoch": 5.18,
+      "eval_accuracy": 0.9971478602318856,
+      "eval_f1": 0.9807878369039392,
+      "eval_loss": 0.014275978319346905,
+      "eval_precision": 0.9776798015982364,
+      "eval_recall": 0.9839156960621187,
+      "eval_runtime": 45.7198,
+      "eval_samples_per_second": 9.383,
+      "eval_steps_per_second": 1.575,
+      "step": 1550
+    },
+    {
+      "epoch": 5.22,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9792703150912107,
+      "eval_loss": 0.01468308549374342,
+      "eval_precision": 0.9760330578512396,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.5952,
+      "eval_samples_per_second": 9.409,
+      "eval_steps_per_second": 1.579,
+      "step": 1560
+    },
+    {
+      "epoch": 5.25,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9810695039380959,
+      "eval_loss": 0.014422405511140823,
+      "eval_precision": 0.977692095841366,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.732,
+      "eval_samples_per_second": 9.381,
+      "eval_steps_per_second": 1.574,
+      "step": 1570
+    },
+    {
+      "epoch": 5.28,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9809075816270061,
+      "eval_loss": 0.01434118952602148,
+      "eval_precision": 0.9787410270568746,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.5148,
+      "eval_samples_per_second": 9.425,
+      "eval_steps_per_second": 1.582,
+      "step": 1580
+    },
+    {
+      "epoch": 5.32,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9791350006908941,
+      "eval_loss": 0.014856048859655857,
+      "eval_precision": 0.9757642522721014,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 46.2893,
+      "eval_samples_per_second": 9.268,
+      "eval_steps_per_second": 1.555,
+      "step": 1590
+    },
+    {
+      "epoch": 5.35,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9803541781959049,
+      "eval_loss": 0.015461008064448833,
+      "eval_precision": 0.9781888459414688,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.2253,
+      "eval_samples_per_second": 9.486,
+      "eval_steps_per_second": 1.592,
+      "step": 1600
+    },
+    {
+      "epoch": 5.38,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9800884955752212,
+      "eval_loss": 0.01574764773249626,
+      "eval_precision": 0.9773855488141202,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.056,
+      "eval_samples_per_second": 9.521,
+      "eval_steps_per_second": 1.598,
+      "step": 1610
+    },
+    {
+      "epoch": 5.42,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9792645839093173,
+      "eval_loss": 0.015540325082838535,
+      "eval_precision": 0.976295479603087,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.1752,
+      "eval_samples_per_second": 9.496,
+      "eval_steps_per_second": 1.594,
+      "step": 1620
+    },
+    {
+      "epoch": 5.45,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9781526548672567,
+      "eval_loss": 0.015259744599461555,
+      "eval_precision": 0.9754550468836183,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.477,
+      "eval_samples_per_second": 9.433,
+      "eval_steps_per_second": 1.583,
+      "step": 1630
+    },
+    {
+      "epoch": 5.48,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9809075816270061,
+      "eval_loss": 0.01531192846596241,
+      "eval_precision": 0.9787410270568746,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.3307,
+      "eval_samples_per_second": 9.464,
+      "eval_steps_per_second": 1.588,
+      "step": 1640
+    },
+    {
+      "epoch": 5.52,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9804790253357331,
+      "eval_loss": 0.015362209640443325,
+      "eval_precision": 0.9789881116947746,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.5289,
+      "eval_samples_per_second": 9.423,
+      "eval_steps_per_second": 1.581,
+      "step": 1650
+    },
+    {
+      "epoch": 5.55,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9736151402127364,
+      "eval_loss": 0.016267232596874237,
+      "eval_precision": 0.9699972474538948,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.4269,
+      "eval_samples_per_second": 9.444,
+      "eval_steps_per_second": 1.585,
+      "step": 1660
+    },
+    {
+      "epoch": 5.59,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9776181265542968,
+      "eval_loss": 0.01579289324581623,
+      "eval_precision": 0.9741189427312775,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 46.1943,
+      "eval_samples_per_second": 9.287,
+      "eval_steps_per_second": 1.559,
+      "step": 1670
+    },
+    {
+      "epoch": 5.62,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9785407725321887,
+      "eval_loss": 0.01532789133489132,
+      "eval_precision": 0.9770528061929776,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 46.2745,
+      "eval_samples_per_second": 9.271,
+      "eval_steps_per_second": 1.556,
+      "step": 1680
+    },
+    {
+      "epoch": 5.65,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.979665237238899,
+      "eval_loss": 0.01536885742098093,
+      "eval_precision": 0.9773668230747999,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.136,
+      "eval_samples_per_second": 9.505,
+      "eval_steps_per_second": 1.595,
+      "step": 1690
+    },
+    {
+      "epoch": 5.69,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9772445179975174,
+      "eval_loss": 0.01578596606850624,
+      "eval_precision": 0.9720164609053498,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.0296,
+      "eval_samples_per_second": 9.527,
+      "eval_steps_per_second": 1.599,
+      "step": 1700
+    },
+    {
+      "epoch": 5.72,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9757040309221425,
+      "eval_loss": 0.01535722240805626,
+      "eval_precision": 0.971412864211105,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.2636,
+      "eval_samples_per_second": 9.478,
+      "eval_steps_per_second": 1.591,
+      "step": 1710
+    },
+    {
+      "epoch": 5.75,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9769496204278812,
+      "eval_loss": 0.01525976974517107,
+      "eval_precision": 0.9725199230557846,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.5287,
+      "eval_samples_per_second": 9.423,
+      "eval_steps_per_second": 1.581,
+      "step": 1720
+    },
+    {
+      "epoch": 5.79,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9776181265542968,
+      "eval_loss": 0.014473304152488708,
+      "eval_precision": 0.9741189427312775,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.6773,
+      "eval_samples_per_second": 9.392,
+      "eval_steps_per_second": 1.576,
+      "step": 1730
+    },
+    {
+      "epoch": 5.82,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9802349689011749,
+      "eval_loss": 0.013995842076838017,
+      "eval_precision": 0.9771286855883163,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.5912,
+      "eval_samples_per_second": 9.41,
+      "eval_steps_per_second": 1.579,
+      "step": 1740
+    },
+    {
+      "epoch": 5.85,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9808864265927978,
+      "eval_loss": 0.014489100314676762,
+      "eval_precision": 0.9798007747648035,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.3576,
+      "eval_samples_per_second": 9.458,
+      "eval_steps_per_second": 1.587,
+      "step": 1750
+    },
+    {
+      "epoch": 5.89,
+      "eval_accuracy": 0.9972274548300656,
+      "eval_f1": 0.9808864265927978,
+      "eval_loss": 0.014342778362333775,
+      "eval_precision": 0.9798007747648035,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.5077,
+      "eval_samples_per_second": 9.427,
+      "eval_steps_per_second": 1.582,
+      "step": 1760
+    },
+    {
+      "epoch": 5.92,
+      "eval_accuracy": 0.9972805178955189,
+      "eval_f1": 0.9800774764803543,
+      "eval_loss": 0.014782003127038479,
+      "eval_precision": 0.9779127553837659,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.3932,
+      "eval_samples_per_second": 9.451,
+      "eval_steps_per_second": 1.586,
+      "step": 1770
+    },
+    {
+      "epoch": 5.95,
+      "eval_accuracy": 0.9972009232973389,
+      "eval_f1": 0.978852798894264,
+      "eval_loss": 0.014779850840568542,
+      "eval_precision": 0.9757508955635161,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.6592,
+      "eval_samples_per_second": 9.396,
+      "eval_steps_per_second": 1.577,
+      "step": 1780
+    },
+    {
+      "epoch": 5.99,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9799141155284665,
+      "eval_loss": 0.015209286473691463,
+      "eval_precision": 0.9789648491558262,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 46.4662,
+      "eval_samples_per_second": 9.233,
+      "eval_steps_per_second": 1.55,
+      "step": 1790
+    },
+    {
+      "epoch": 6.02,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.981011781011781,
+      "eval_loss": 0.015130845829844475,
+      "eval_precision": 0.9806040454419507,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.3314,
+      "eval_samples_per_second": 9.464,
+      "eval_steps_per_second": 1.588,
+      "step": 1800
+    },
+    {
+      "epoch": 6.05,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9795297372060858,
+      "eval_loss": 0.015158111229538918,
+      "eval_precision": 0.9770971302428256,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.4446,
+      "eval_samples_per_second": 9.44,
+      "eval_steps_per_second": 1.584,
+      "step": 1810
+    },
+    {
+      "epoch": 6.09,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9782638792745397,
+      "eval_loss": 0.015132046304643154,
+      "eval_precision": 0.9767763339784352,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.4433,
+      "eval_samples_per_second": 9.44,
+      "eval_steps_per_second": 1.584,
+      "step": 1820
+    },
+    {
+      "epoch": 6.12,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9705516383243468,
+      "eval_loss": 0.01593286544084549,
+      "eval_precision": 0.967741935483871,
+      "eval_recall": 0.9733777038269551,
+      "eval_runtime": 45.4009,
+      "eval_samples_per_second": 9.449,
+      "eval_steps_per_second": 1.586,
+      "step": 1830
+    },
+    {
+      "epoch": 6.15,
+      "eval_accuracy": 0.9969356079700724,
+      "eval_f1": 0.9730774540936076,
+      "eval_loss": 0.014840428717434406,
+      "eval_precision": 0.9689304371734946,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.3952,
+      "eval_samples_per_second": 9.45,
+      "eval_steps_per_second": 1.586,
+      "step": 1840
+    },
+    {
+      "epoch": 6.19,
+      "eval_accuracy": 0.9966570268764426,
+      "eval_f1": 0.9764868603042877,
+      "eval_loss": 0.016662631183862686,
+      "eval_precision": 0.9740618101545254,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.439,
+      "eval_samples_per_second": 9.441,
+      "eval_steps_per_second": 1.585,
+      "step": 1850
+    },
+    {
+      "epoch": 6.22,
+      "eval_accuracy": 0.9965376349791727,
+      "eval_f1": 0.9750761561894212,
+      "eval_loss": 0.01773080602288246,
+      "eval_precision": 0.9737278761061947,
+      "eval_recall": 0.9764281752634498,
+      "eval_runtime": 45.6526,
+      "eval_samples_per_second": 9.397,
+      "eval_steps_per_second": 1.577,
+      "step": 1860
+    },
+    {
+      "epoch": 6.25,
+      "eval_accuracy": 0.9964182430819029,
+      "eval_f1": 0.9697221070095395,
+      "eval_loss": 0.018397442996501923,
+      "eval_precision": 0.9669148056244831,
+      "eval_recall": 0.9725457570715474,
+      "eval_runtime": 45.4014,
+      "eval_samples_per_second": 9.449,
+      "eval_steps_per_second": 1.586,
+      "step": 1870
+    },
+    {
+      "epoch": 6.29,
+      "eval_accuracy": 0.9966968241755326,
+      "eval_f1": 0.9773292784075199,
+      "eval_loss": 0.0170457661151886,
+      "eval_precision": 0.9743660418963617,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.6124,
+      "eval_samples_per_second": 9.405,
+      "eval_steps_per_second": 1.579,
+      "step": 1880
+    },
+    {
+      "epoch": 6.32,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9793046357615895,
+      "eval_loss": 0.01652824692428112,
+      "eval_precision": 0.9744645799011532,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.4345,
+      "eval_samples_per_second": 9.442,
+      "eval_steps_per_second": 1.585,
+      "step": 1890
+    },
+    {
+      "epoch": 6.35,
+      "eval_accuracy": 0.9967233557082593,
+      "eval_f1": 0.978476821192053,
+      "eval_loss": 0.016787514090538025,
+      "eval_precision": 0.9736408566721582,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.5818,
+      "eval_samples_per_second": 9.412,
+      "eval_steps_per_second": 1.58,
+      "step": 1900
+    },
+    {
+      "epoch": 6.39,
+      "eval_accuracy": 0.9965641665118994,
+      "eval_f1": 0.975320557010892,
+      "eval_loss": 0.01782037876546383,
+      "eval_precision": 0.9698382231971483,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.9556,
+      "eval_samples_per_second": 9.335,
+      "eval_steps_per_second": 1.567,
+      "step": 1910
+    },
+    {
+      "epoch": 6.42,
+      "eval_accuracy": 0.9966702926428059,
+      "eval_f1": 0.9768211920529801,
+      "eval_loss": 0.017588861286640167,
+      "eval_precision": 0.9719934102141681,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.1589,
+      "eval_samples_per_second": 9.5,
+      "eval_steps_per_second": 1.594,
+      "step": 1920
+    },
+    {
+      "epoch": 6.45,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9810799613313078,
+      "eval_loss": 0.016702750697731972,
+      "eval_precision": 0.9771664374140303,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.117,
+      "eval_samples_per_second": 9.509,
+      "eval_steps_per_second": 1.596,
+      "step": 1930
+    },
+    {
+      "epoch": 6.49,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9783179118906229,
+      "eval_loss": 0.016243569552898407,
+      "eval_precision": 0.9744154057771665,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.1078,
+      "eval_samples_per_second": 9.511,
+      "eval_steps_per_second": 1.596,
+      "step": 1940
+    },
+    {
+      "epoch": 6.52,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9773418071290412,
+      "eval_loss": 0.01623663119971752,
+      "eval_precision": 0.9738436123348018,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.7949,
+      "eval_samples_per_second": 9.368,
+      "eval_steps_per_second": 1.572,
+      "step": 1950
+    },
+    {
+      "epoch": 6.56,
+      "eval_accuracy": 0.9967896845400759,
+      "eval_f1": 0.9769496204278812,
+      "eval_loss": 0.016976099461317062,
+      "eval_precision": 0.9725199230557846,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.3661,
+      "eval_samples_per_second": 9.456,
+      "eval_steps_per_second": 1.587,
+      "step": 1960
+    },
+    {
+      "epoch": 6.59,
+      "eval_accuracy": 0.9967233557082593,
+      "eval_f1": 0.9759933774834438,
+      "eval_loss": 0.017524730414152145,
+      "eval_precision": 0.971169686985173,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.7337,
+      "eval_samples_per_second": 9.38,
+      "eval_steps_per_second": 1.574,
+      "step": 1970
+    },
+    {
+      "epoch": 6.62,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9794397681799365,
+      "eval_loss": 0.016901282593607903,
+      "eval_precision": 0.9747322164240593,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.4807,
+      "eval_samples_per_second": 9.433,
+      "eval_steps_per_second": 1.583,
+      "step": 1980
+    },
+    {
+      "epoch": 6.66,
+      "eval_accuracy": 0.9967896845400759,
+      "eval_f1": 0.9788761562888305,
+      "eval_loss": 0.01667814515531063,
+      "eval_precision": 0.9747044267253231,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.6668,
+      "eval_samples_per_second": 9.394,
+      "eval_steps_per_second": 1.577,
+      "step": 1990
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.0064,
+      "step": 2000
+    },
+    {
+      "epoch": 6.69,
+      "eval_accuracy": 0.9967100899418959,
+      "eval_f1": 0.9776428374275461,
+      "eval_loss": 0.016720809042453766,
+      "eval_precision": 0.9730769230769231,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.4912,
+      "eval_samples_per_second": 9.43,
+      "eval_steps_per_second": 1.583,
+      "step": 2000
+    },
+    {
+      "epoch": 6.72,
+      "eval_accuracy": 0.9966172295773527,
+      "eval_f1": 0.9770781552057443,
+      "eval_loss": 0.017004678025841713,
+      "eval_precision": 0.9730473047304731,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.7222,
+      "eval_samples_per_second": 9.383,
+      "eval_steps_per_second": 1.575,
+      "step": 2010
+    },
+    {
+      "epoch": 6.76,
+      "eval_accuracy": 0.9966172295773527,
+      "eval_f1": 0.9774892970584174,
+      "eval_loss": 0.016945617273449898,
+      "eval_precision": 0.9735900962861073,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.3731,
+      "eval_samples_per_second": 9.455,
+      "eval_steps_per_second": 1.587,
+      "step": 2020
+    },
+    {
+      "epoch": 6.79,
+      "eval_accuracy": 0.9967100899418959,
+      "eval_f1": 0.9788994621431527,
+      "eval_loss": 0.016928063705563545,
+      "eval_precision": 0.9736625514403292,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.8513,
+      "eval_samples_per_second": 9.356,
+      "eval_steps_per_second": 1.57,
+      "step": 2030
+    },
+    {
+      "epoch": 6.82,
+      "eval_accuracy": 0.9967896845400759,
+      "eval_f1": 0.9779675020655467,
+      "eval_loss": 0.01666436158120632,
+      "eval_precision": 0.9712800875273523,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.4258,
+      "eval_samples_per_second": 9.444,
+      "eval_steps_per_second": 1.585,
+      "step": 2040
+    },
+    {
+      "epoch": 6.86,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9770130763936682,
+      "eval_loss": 0.016234010457992554,
+      "eval_precision": 0.969937141295436,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.6745,
+      "eval_samples_per_second": 9.393,
+      "eval_steps_per_second": 1.576,
+      "step": 2050
+    },
+    {
+      "epoch": 6.89,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9758986365514392,
+      "eval_loss": 0.015693064779043198,
+      "eval_precision": 0.9693570451436389,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 46.507,
+      "eval_samples_per_second": 9.224,
+      "eval_steps_per_second": 1.548,
+      "step": 2060
+    },
+    {
+      "epoch": 6.92,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.978476821192053,
+      "eval_loss": 0.015247814357280731,
+      "eval_precision": 0.9736408566721582,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.4327,
+      "eval_samples_per_second": 9.443,
+      "eval_steps_per_second": 1.585,
+      "step": 2070
+    },
+    {
+      "epoch": 6.96,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9795636564484949,
+      "eval_loss": 0.01587117277085781,
+      "eval_precision": 0.9755225522552256,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.1551,
+      "eval_samples_per_second": 9.501,
+      "eval_steps_per_second": 1.595,
+      "step": 2080
+    },
+    {
+      "epoch": 6.99,
+      "eval_accuracy": 0.9967100899418959,
+      "eval_f1": 0.9783776339347198,
+      "eval_loss": 0.017061389982700348,
+      "eval_precision": 0.9718194254445964,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.1957,
+      "eval_samples_per_second": 9.492,
+      "eval_steps_per_second": 1.593,
+      "step": 2090
+    },
+    {
+      "epoch": 7.02,
+      "eval_accuracy": 0.9965906980446261,
+      "eval_f1": 0.9775821757667446,
+      "eval_loss": 0.018233157694339752,
+      "eval_precision": 0.9697135061391542,
+      "eval_recall": 0.985579589572934,
+      "eval_runtime": 45.2503,
+      "eval_samples_per_second": 9.481,
+      "eval_steps_per_second": 1.591,
+      "step": 2100
+    },
+    {
+      "epoch": 7.06,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9790113228389947,
+      "eval_loss": 0.01590120792388916,
+      "eval_precision": 0.974972497249725,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.4565,
+      "eval_samples_per_second": 9.438,
+      "eval_steps_per_second": 1.584,
+      "step": 2110
+    },
+    {
+      "epoch": 7.09,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9770971302428255,
+      "eval_loss": 0.015739668160676956,
+      "eval_precision": 0.972267984623833,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 46.0012,
+      "eval_samples_per_second": 9.326,
+      "eval_steps_per_second": 1.565,
+      "step": 2120
+    },
+    {
+      "epoch": 7.12,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9790113228389947,
+      "eval_loss": 0.015736253932118416,
+      "eval_precision": 0.974972497249725,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.4255,
+      "eval_samples_per_second": 9.444,
+      "eval_steps_per_second": 1.585,
+      "step": 2130
+    },
+    {
+      "epoch": 7.16,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9784827586206897,
+      "eval_loss": 0.016040045768022537,
+      "eval_precision": 0.9733809001097695,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.5296,
+      "eval_samples_per_second": 9.422,
+      "eval_steps_per_second": 1.581,
+      "step": 2140
+    },
+    {
+      "epoch": 7.19,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9786413118368471,
+      "eval_loss": 0.01620173640549183,
+      "eval_precision": 0.9726102437688304,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.6034,
+      "eval_samples_per_second": 9.407,
+      "eval_steps_per_second": 1.579,
+      "step": 2150
+    },
+    {
+      "epoch": 7.22,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9798565121412803,
+      "eval_loss": 0.0158588457852602,
+      "eval_precision": 0.9750137287204832,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.789,
+      "eval_samples_per_second": 9.369,
+      "eval_steps_per_second": 1.572,
+      "step": 2160
+    },
+    {
+      "epoch": 7.26,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9791522849647936,
+      "eval_loss": 0.015738798305392265,
+      "eval_precision": 0.9749793786087435,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.511,
+      "eval_samples_per_second": 9.426,
+      "eval_steps_per_second": 1.582,
+      "step": 2170
+    },
+    {
+      "epoch": 7.29,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9794113582976371,
+      "eval_loss": 0.015642931684851646,
+      "eval_precision": 0.976039658496282,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.1132,
+      "eval_samples_per_second": 9.509,
+      "eval_steps_per_second": 1.596,
+      "step": 2180
+    },
+    {
+      "epoch": 7.32,
+      "eval_accuracy": 0.9971611259982489,
+      "eval_f1": 0.9811998894111142,
+      "eval_loss": 0.015118683688342571,
+      "eval_precision": 0.9782249173098125,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.0629,
+      "eval_samples_per_second": 9.52,
+      "eval_steps_per_second": 1.598,
+      "step": 2190
+    },
+    {
+      "epoch": 7.36,
+      "eval_accuracy": 0.9973601124936987,
+      "eval_f1": 0.982160143825197,
+      "eval_loss": 0.015025600790977478,
+      "eval_precision": 0.9795862068965517,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.3715,
+      "eval_samples_per_second": 9.455,
+      "eval_steps_per_second": 1.587,
+      "step": 2200
+    },
+    {
+      "epoch": 7.39,
+      "eval_accuracy": 0.9972539863627922,
+      "eval_f1": 0.9786821705426357,
+      "eval_loss": 0.014340460300445557,
+      "eval_precision": 0.9770591487009398,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.1287,
+      "eval_samples_per_second": 9.506,
+      "eval_steps_per_second": 1.595,
+      "step": 2210
+    },
+    {
+      "epoch": 7.42,
+      "eval_accuracy": 0.9973335809609721,
+      "eval_f1": 0.9813097051086805,
+      "eval_loss": 0.0135923121124506,
+      "eval_precision": 0.979817528338402,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.1943,
+      "eval_samples_per_second": 9.492,
+      "eval_steps_per_second": 1.593,
+      "step": 2220
+    },
+    {
+      "epoch": 7.46,
+      "eval_accuracy": 0.9972009232973389,
+      "eval_f1": 0.978570441034149,
+      "eval_loss": 0.014164491556584835,
+      "eval_precision": 0.9757375241246209,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.309,
+      "eval_samples_per_second": 9.468,
+      "eval_steps_per_second": 1.589,
+      "step": 2230
+    },
+    {
+      "epoch": 7.49,
+      "eval_accuracy": 0.9972672521291555,
+      "eval_f1": 0.9806308799114555,
+      "eval_loss": 0.01402602344751358,
+      "eval_precision": 0.9784649364991718,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.5204,
+      "eval_samples_per_second": 9.424,
+      "eval_steps_per_second": 1.582,
+      "step": 2240
+    },
+    {
+      "epoch": 7.53,
+      "eval_accuracy": 0.9972539863627922,
+      "eval_f1": 0.9809181415929205,
+      "eval_loss": 0.014084520749747753,
+      "eval_precision": 0.9782129067843354,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.1882,
+      "eval_samples_per_second": 9.494,
+      "eval_steps_per_second": 1.593,
+      "step": 2250
+    },
+    {
+      "epoch": 7.56,
+      "eval_accuracy": 0.9972937836618821,
+      "eval_f1": 0.9835111542192047,
+      "eval_loss": 0.014140544459223747,
+      "eval_precision": 0.9828302409304902,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.1162,
+      "eval_samples_per_second": 9.509,
+      "eval_steps_per_second": 1.596,
+      "step": 2260
+    },
+    {
+      "epoch": 7.59,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.980484429065744,
+      "eval_loss": 0.015128599479794502,
+      "eval_precision": 0.9787234042553191,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.1622,
+      "eval_samples_per_second": 9.499,
+      "eval_steps_per_second": 1.594,
+      "step": 2270
+    },
+    {
+      "epoch": 7.63,
+      "eval_accuracy": 0.9971743917646122,
+      "eval_f1": 0.9838017444275232,
+      "eval_loss": 0.014916467480361462,
+      "eval_precision": 0.9823057782692839,
+      "eval_recall": 0.9853022739877981,
+      "eval_runtime": 45.3078,
+      "eval_samples_per_second": 9.469,
+      "eval_steps_per_second": 1.589,
+      "step": 2280
+    },
+    {
+      "epoch": 7.66,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.9826989619377162,
+      "eval_loss": 0.014806441031396389,
+      "eval_precision": 0.980933959657364,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.215,
+      "eval_samples_per_second": 9.488,
+      "eval_steps_per_second": 1.592,
+      "step": 2290
+    },
+    {
+      "epoch": 7.69,
+      "eval_accuracy": 0.9973335809609721,
+      "eval_f1": 0.9828444936358606,
+      "eval_loss": 0.014511052519083023,
+      "eval_precision": 0.9806736609607951,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.1721,
+      "eval_samples_per_second": 9.497,
+      "eval_steps_per_second": 1.594,
+      "step": 2300
+    },
+    {
+      "epoch": 7.73,
+      "eval_accuracy": 0.9973070494282454,
+      "eval_f1": 0.9824124082537044,
+      "eval_loss": 0.014819289557635784,
+      "eval_precision": 0.9811894882434301,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.484,
+      "eval_samples_per_second": 9.432,
+      "eval_steps_per_second": 1.583,
+      "step": 2310
+    },
+    {
+      "epoch": 7.76,
+      "eval_accuracy": 0.9971611259982489,
+      "eval_f1": 0.9818935729094679,
+      "eval_loss": 0.015233664773404598,
+      "eval_precision": 0.9787820336180766,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.245,
+      "eval_samples_per_second": 9.482,
+      "eval_steps_per_second": 1.591,
+      "step": 2320
+    },
+    {
+      "epoch": 7.79,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9814865985078752,
+      "eval_loss": 0.01592710055410862,
+      "eval_precision": 0.9779735682819384,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.1068,
+      "eval_samples_per_second": 9.511,
+      "eval_steps_per_second": 1.596,
+      "step": 2330
+    },
+    {
+      "epoch": 7.83,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9807984528249758,
+      "eval_loss": 0.01635568030178547,
+      "eval_precision": 0.9771538673272777,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.3054,
+      "eval_samples_per_second": 9.469,
+      "eval_steps_per_second": 1.589,
+      "step": 2340
+    },
+    {
+      "epoch": 7.86,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9810747340792927,
+      "eval_loss": 0.015507887117564678,
+      "eval_precision": 0.9774291219377924,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.3049,
+      "eval_samples_per_second": 9.469,
+      "eval_steps_per_second": 1.589,
+      "step": 2350
+    },
+    {
+      "epoch": 7.89,
+      "eval_accuracy": 0.9971743917646122,
+      "eval_f1": 0.9814814814814814,
+      "eval_loss": 0.01488608680665493,
+      "eval_precision": 0.978236914600551,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.3669,
+      "eval_samples_per_second": 9.456,
+      "eval_steps_per_second": 1.587,
+      "step": 2360
+    },
+    {
+      "epoch": 7.93,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9790055248618784,
+      "eval_loss": 0.016140291467308998,
+      "eval_precision": 0.9752339020363237,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.5755,
+      "eval_samples_per_second": 9.413,
+      "eval_steps_per_second": 1.58,
+      "step": 2370
+    },
+    {
+      "epoch": 7.96,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.980511402902557,
+      "eval_loss": 0.01678432524204254,
+      "eval_precision": 0.9774042435932764,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.3412,
+      "eval_samples_per_second": 9.462,
+      "eval_steps_per_second": 1.588,
+      "step": 2380
+    },
+    {
+      "epoch": 7.99,
+      "eval_accuracy": 0.9968162160728026,
+      "eval_f1": 0.9803704727674868,
+      "eval_loss": 0.016629330813884735,
+      "eval_precision": 0.9773980154355016,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.5099,
+      "eval_samples_per_second": 9.427,
+      "eval_steps_per_second": 1.582,
+      "step": 2390
+    },
+    {
+      "epoch": 8.03,
+      "eval_accuracy": 0.9968560133718924,
+      "eval_f1": 0.9798119469026549,
+      "eval_loss": 0.01662250980734825,
+      "eval_precision": 0.9771097628240485,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.337,
+      "eval_samples_per_second": 9.462,
+      "eval_steps_per_second": 1.588,
+      "step": 2400
+    },
+    {
+      "epoch": 8.06,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9813355454168395,
+      "eval_loss": 0.0163425225764513,
+      "eval_precision": 0.978494623655914,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.4794,
+      "eval_samples_per_second": 9.433,
+      "eval_steps_per_second": 1.583,
+      "step": 2410
+    },
+    {
+      "epoch": 8.09,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9811842833425568,
+      "eval_loss": 0.016381612047553062,
+      "eval_precision": 0.9790171176145775,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.1726,
+      "eval_samples_per_second": 9.497,
+      "eval_steps_per_second": 1.594,
+      "step": 2420
+    },
+    {
+      "epoch": 8.13,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9817528338401991,
+      "eval_loss": 0.016340401023626328,
+      "eval_precision": 0.9787761852260198,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.3524,
+      "eval_samples_per_second": 9.459,
+      "eval_steps_per_second": 1.588,
+      "step": 2430
+    },
+    {
+      "epoch": 8.16,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.980771890994605,
+      "eval_loss": 0.016326354816555977,
+      "eval_precision": 0.9784708804857852,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.2927,
+      "eval_samples_per_second": 9.472,
+      "eval_steps_per_second": 1.59,
+      "step": 2440
+    },
+    {
+      "epoch": 8.19,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9817275747508305,
+      "eval_loss": 0.015510362572968006,
+      "eval_precision": 0.9800995024875622,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.5096,
+      "eval_samples_per_second": 9.427,
+      "eval_steps_per_second": 1.582,
+      "step": 2450
+    },
+    {
+      "epoch": 8.23,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9824464409122322,
+      "eval_loss": 0.015236412174999714,
+      "eval_precision": 0.9793331496279967,
+      "eval_recall": 0.985579589572934,
+      "eval_runtime": 45.3704,
+      "eval_samples_per_second": 9.455,
+      "eval_steps_per_second": 1.587,
+      "step": 2460
+    },
+    {
+      "epoch": 8.26,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9814865985078752,
+      "eval_loss": 0.015579747967422009,
+      "eval_precision": 0.9779735682819384,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.2219,
+      "eval_samples_per_second": 9.487,
+      "eval_steps_per_second": 1.592,
+      "step": 2470
+    },
+    {
+      "epoch": 8.29,
+      "eval_accuracy": 0.9972009232973389,
+      "eval_f1": 0.9789590254706533,
+      "eval_loss": 0.014976629987359047,
+      "eval_precision": 0.9773355444997236,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.5521,
+      "eval_samples_per_second": 9.418,
+      "eval_steps_per_second": 1.581,
+      "step": 2480
+    },
+    {
+      "epoch": 8.33,
+      "eval_accuracy": 0.9973335809609721,
+      "eval_f1": 0.9824318716281643,
+      "eval_loss": 0.014180959202349186,
+      "eval_precision": 0.9801269666022633,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.6388,
+      "eval_samples_per_second": 9.4,
+      "eval_steps_per_second": 1.578,
+      "step": 2490
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.0051,
+      "step": 2500
+    },
+    {
+      "epoch": 8.36,
+      "eval_accuracy": 0.9972274548300656,
+      "eval_f1": 0.9791522849647936,
+      "eval_loss": 0.014656171202659607,
+      "eval_precision": 0.9749793786087435,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.2727,
+      "eval_samples_per_second": 9.476,
+      "eval_steps_per_second": 1.59,
+      "step": 2500
+    },
+    {
+      "epoch": 8.39,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9780477702609416,
+      "eval_loss": 0.014881155453622341,
+      "eval_precision": 0.9738795710750618,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 45.7124,
+      "eval_samples_per_second": 9.385,
+      "eval_steps_per_second": 1.575,
+      "step": 2510
+    },
+    {
+      "epoch": 8.43,
+      "eval_accuracy": 0.9972141890637022,
+      "eval_f1": 0.9806629834254144,
+      "eval_loss": 0.014929546974599361,
+      "eval_precision": 0.976884975233902,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.3886,
+      "eval_samples_per_second": 9.452,
+      "eval_steps_per_second": 1.586,
+      "step": 2520
+    },
+    {
+      "epoch": 8.46,
+      "eval_accuracy": 0.9972009232973389,
+      "eval_f1": 0.9802404311178665,
+      "eval_loss": 0.014953644014894962,
+      "eval_precision": 0.976865877168824,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.8505,
+      "eval_samples_per_second": 9.356,
+      "eval_steps_per_second": 1.57,
+      "step": 2530
+    },
+    {
+      "epoch": 8.49,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9813252178724583,
+      "eval_loss": 0.014914426021277905,
+      "eval_precision": 0.979022909191278,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.5694,
+      "eval_samples_per_second": 9.414,
+      "eval_steps_per_second": 1.58,
+      "step": 2540
+    },
+    {
+      "epoch": 8.53,
+      "eval_accuracy": 0.9972672521291555,
+      "eval_f1": 0.9826941713969264,
+      "eval_loss": 0.014766544103622437,
+      "eval_precision": 0.9811998894111141,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.2509,
+      "eval_samples_per_second": 9.48,
+      "eval_steps_per_second": 1.591,
+      "step": 2550
+    },
+    {
+      "epoch": 8.56,
+      "eval_accuracy": 0.9972407205964289,
+      "eval_f1": 0.9839468585662884,
+      "eval_loss": 0.015172240324318409,
+      "eval_precision": 0.9820441988950276,
+      "eval_recall": 0.9858569051580699,
+      "eval_runtime": 45.7001,
+      "eval_samples_per_second": 9.387,
+      "eval_steps_per_second": 1.575,
+      "step": 2560
+    },
+    {
+      "epoch": 8.6,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9829993089149965,
+      "eval_loss": 0.015841683372855186,
+      "eval_precision": 0.9798842656379168,
+      "eval_recall": 0.9861342207432058,
+      "eval_runtime": 45.6884,
+      "eval_samples_per_second": 9.39,
+      "eval_steps_per_second": 1.576,
+      "step": 2570
+    },
+    {
+      "epoch": 8.63,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9813407049067036,
+      "eval_loss": 0.016232503578066826,
+      "eval_precision": 0.9782309176081565,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.3554,
+      "eval_samples_per_second": 9.459,
+      "eval_steps_per_second": 1.587,
+      "step": 2580
+    },
+    {
+      "epoch": 8.66,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9763714246234628,
+      "eval_loss": 0.015989486128091812,
+      "eval_precision": 0.9730101900302947,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.3338,
+      "eval_samples_per_second": 9.463,
+      "eval_steps_per_second": 1.588,
+      "step": 2590
+    },
+    {
+      "epoch": 8.7,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9746922970543493,
+      "eval_loss": 0.015164612792432308,
+      "eval_precision": 0.9721379310344828,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.3618,
+      "eval_samples_per_second": 9.457,
+      "eval_steps_per_second": 1.587,
+      "step": 2600
+    },
+    {
+      "epoch": 8.73,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9752111895859299,
+      "eval_loss": 0.015361527912318707,
+      "eval_precision": 0.9739972337482711,
+      "eval_recall": 0.9764281752634498,
+      "eval_runtime": 45.7043,
+      "eval_samples_per_second": 9.386,
+      "eval_steps_per_second": 1.575,
+      "step": 2610
+    },
+    {
+      "epoch": 8.76,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9776057506220625,
+      "eval_loss": 0.015215002000331879,
+      "eval_precision": 0.9746416758544653,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.5622,
+      "eval_samples_per_second": 9.416,
+      "eval_steps_per_second": 1.58,
+      "step": 2620
+    },
+    {
+      "epoch": 8.8,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9777470628887354,
+      "eval_loss": 0.014545532874763012,
+      "eval_precision": 0.974648663543676,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.4027,
+      "eval_samples_per_second": 9.449,
+      "eval_steps_per_second": 1.586,
+      "step": 2630
+    },
+    {
+      "epoch": 8.83,
+      "eval_accuracy": 0.9971611259982489,
+      "eval_f1": 0.9813355454168395,
+      "eval_loss": 0.014298198744654655,
+      "eval_precision": 0.978494623655914,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.6015,
+      "eval_samples_per_second": 9.408,
+      "eval_steps_per_second": 1.579,
+      "step": 2640
+    },
+    {
+      "epoch": 8.86,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9804952275556785,
+      "eval_loss": 0.014669723808765411,
+      "eval_precision": 0.978194866133039,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.6583,
+      "eval_samples_per_second": 9.396,
+      "eval_steps_per_second": 1.577,
+      "step": 2650
+    },
+    {
+      "epoch": 8.9,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9749550297495504,
+      "eval_loss": 0.014710523188114166,
+      "eval_precision": 0.9729356531344933,
+      "eval_recall": 0.9769828064337216,
+      "eval_runtime": 45.2863,
+      "eval_samples_per_second": 9.473,
+      "eval_steps_per_second": 1.59,
+      "step": 2660
+    },
+    {
+      "epoch": 8.93,
+      "eval_accuracy": 0.9972141890637022,
+      "eval_f1": 0.9820094104622197,
+      "eval_loss": 0.014592879451811314,
+      "eval_precision": 0.980110497237569,
+      "eval_recall": 0.9839156960621187,
+      "eval_runtime": 45.3618,
+      "eval_samples_per_second": 9.457,
+      "eval_steps_per_second": 1.587,
+      "step": 2670
+    },
+    {
+      "epoch": 8.96,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.9817477876106194,
+      "eval_loss": 0.014163294807076454,
+      "eval_precision": 0.9790402647545504,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.3659,
+      "eval_samples_per_second": 9.456,
+      "eval_steps_per_second": 1.587,
+      "step": 2680
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9792932081722805,
+      "eval_loss": 0.01418287307024002,
+      "eval_precision": 0.9749862561847169,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.6259,
+      "eval_samples_per_second": 9.403,
+      "eval_steps_per_second": 1.578,
+      "step": 2690
+    },
+    {
+      "epoch": 9.03,
+      "eval_accuracy": 0.9971478602318856,
+      "eval_f1": 0.9809286898839138,
+      "eval_loss": 0.014496715739369392,
+      "eval_precision": 0.9776859504132231,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.371,
+      "eval_samples_per_second": 9.455,
+      "eval_steps_per_second": 1.587,
+      "step": 2700
+    },
+    {
+      "epoch": 9.06,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9802295036637633,
+      "eval_loss": 0.014654590748250484,
+      "eval_precision": 0.9773917838433968,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.386,
+      "eval_samples_per_second": 9.452,
+      "eval_steps_per_second": 1.586,
+      "step": 2710
+    },
+    {
+      "epoch": 9.1,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9795410561238596,
+      "eval_loss": 0.014612732455134392,
+      "eval_precision": 0.9765711135611908,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.3853,
+      "eval_samples_per_second": 9.452,
+      "eval_steps_per_second": 1.586,
+      "step": 2720
+    },
+    {
+      "epoch": 9.13,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9796989366109652,
+      "eval_loss": 0.014642550610005856,
+      "eval_precision": 0.9757909215955983,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.583,
+      "eval_samples_per_second": 9.411,
+      "eval_steps_per_second": 1.58,
+      "step": 2730
+    },
+    {
+      "epoch": 9.16,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9792817679558011,
+      "eval_loss": 0.014756478369235992,
+      "eval_precision": 0.9755090809025867,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.3958,
+      "eval_samples_per_second": 9.45,
+      "eval_steps_per_second": 1.586,
+      "step": 2740
+    },
+    {
+      "epoch": 9.2,
+      "eval_accuracy": 0.9971611259982489,
+      "eval_f1": 0.9781526548672567,
+      "eval_loss": 0.014514540322124958,
+      "eval_precision": 0.9754550468836183,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.3479,
+      "eval_samples_per_second": 9.46,
+      "eval_steps_per_second": 1.588,
+      "step": 2750
+    },
+    {
+      "epoch": 9.23,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.9780234968901176,
+      "eval_loss": 0.014607676304876804,
+      "eval_precision": 0.974924221548636,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.3679,
+      "eval_samples_per_second": 9.456,
+      "eval_steps_per_second": 1.587,
+      "step": 2760
+    },
+    {
+      "epoch": 9.26,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9799585348997927,
+      "eval_loss": 0.014230134896934032,
+      "eval_precision": 0.9768531275833563,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.6606,
+      "eval_samples_per_second": 9.395,
+      "eval_steps_per_second": 1.577,
+      "step": 2770
+    },
+    {
+      "epoch": 9.3,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9814814814814814,
+      "eval_loss": 0.014004090800881386,
+      "eval_precision": 0.978236914600551,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.4655,
+      "eval_samples_per_second": 9.436,
+      "eval_steps_per_second": 1.584,
+      "step": 2780
+    },
+    {
+      "epoch": 9.33,
+      "eval_accuracy": 0.9971080629327956,
+      "eval_f1": 0.980359612724758,
+      "eval_loss": 0.013850794173777103,
+      "eval_precision": 0.977924944812362,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.4768,
+      "eval_samples_per_second": 9.433,
+      "eval_steps_per_second": 1.583,
+      "step": 2790
+    },
+    {
+      "epoch": 9.36,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9795070617557464,
+      "eval_loss": 0.014602423645555973,
+      "eval_precision": 0.9781526548672567,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 46.0797,
+      "eval_samples_per_second": 9.31,
+      "eval_steps_per_second": 1.563,
+      "step": 2800
+    },
+    {
+      "epoch": 9.4,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9811738648947951,
+      "eval_loss": 0.01487213745713234,
+      "eval_precision": 0.9795467108899945,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.4621,
+      "eval_samples_per_second": 9.436,
+      "eval_steps_per_second": 1.584,
+      "step": 2810
+    },
+    {
+      "epoch": 9.43,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9789648491558263,
+      "eval_loss": 0.01481552142649889,
+      "eval_precision": 0.9770718232044199,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.388,
+      "eval_samples_per_second": 9.452,
+      "eval_steps_per_second": 1.586,
+      "step": 2820
+    },
+    {
+      "epoch": 9.46,
+      "eval_accuracy": 0.9973601124936987,
+      "eval_f1": 0.9822763777346994,
+      "eval_loss": 0.01404637098312378,
+      "eval_precision": 0.9809181415929203,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.4647,
+      "eval_samples_per_second": 9.436,
+      "eval_steps_per_second": 1.584,
+      "step": 2830
+    },
+    {
+      "epoch": 9.5,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9712539925010416,
+      "eval_loss": 0.014825278893113136,
+      "eval_precision": 0.972739916550765,
+      "eval_recall": 0.9697726012201886,
+      "eval_runtime": 45.3762,
+      "eval_samples_per_second": 9.454,
+      "eval_steps_per_second": 1.587,
+      "step": 2840
+    },
+    {
+      "epoch": 9.53,
+      "eval_accuracy": 0.9964978376800828,
+      "eval_f1": 0.9593767390094602,
+      "eval_loss": 0.015539586544036865,
+      "eval_precision": 0.9625907314349526,
+      "eval_recall": 0.9561841375485303,
+      "eval_runtime": 45.6548,
+      "eval_samples_per_second": 9.397,
+      "eval_steps_per_second": 1.577,
+      "step": 2850
+    },
+    {
+      "epoch": 9.57,
+      "eval_accuracy": 0.9964978376800828,
+      "eval_f1": 0.9605006954102921,
+      "eval_loss": 0.016403868794441223,
+      "eval_precision": 0.9634486607142857,
+      "eval_recall": 0.9575707154742097,
+      "eval_runtime": 45.6372,
+      "eval_samples_per_second": 9.4,
+      "eval_steps_per_second": 1.578,
+      "step": 2860
+    },
+    {
+      "epoch": 9.6,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9791350006908941,
+      "eval_loss": 0.015382306650280952,
+      "eval_precision": 0.9757642522721014,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.4505,
+      "eval_samples_per_second": 9.439,
+      "eval_steps_per_second": 1.584,
+      "step": 2870
+    },
+    {
+      "epoch": 9.63,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9781647318960752,
+      "eval_loss": 0.015797466039657593,
+      "eval_precision": 0.974931129476584,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.42,
+      "eval_samples_per_second": 9.445,
+      "eval_steps_per_second": 1.585,
+      "step": 2880
+    },
+    {
+      "epoch": 9.67,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9791637919138954,
+      "eval_loss": 0.01617647521197796,
+      "eval_precision": 0.9744575666025818,
+      "eval_recall": 0.9839156960621187,
+      "eval_runtime": 45.4393,
+      "eval_samples_per_second": 9.441,
+      "eval_steps_per_second": 1.585,
+      "step": 2890
+    },
+    {
+      "epoch": 9.7,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.978629532607197,
+      "eval_loss": 0.0164639949798584,
+      "eval_precision": 0.9731285988483686,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.4559,
+      "eval_samples_per_second": 9.438,
+      "eval_steps_per_second": 1.584,
+      "step": 2900
+    },
+    {
+      "epoch": 9.73,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9791810285399146,
+      "eval_loss": 0.016681063920259476,
+      "eval_precision": 0.9736769947902385,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.5001,
+      "eval_samples_per_second": 9.429,
+      "eval_steps_per_second": 1.582,
+      "step": 2910
+    },
+    {
+      "epoch": 9.77,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9808090570205716,
+      "eval_loss": 0.016886161640286446,
+      "eval_precision": 0.9766290899092659,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.9762,
+      "eval_samples_per_second": 9.331,
+      "eval_steps_per_second": 1.566,
+      "step": 2920
+    },
+    {
+      "epoch": 9.8,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9787351560342448,
+      "eval_loss": 0.017214221879839897,
+      "eval_precision": 0.9746974697469747,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.4576,
+      "eval_samples_per_second": 9.437,
+      "eval_steps_per_second": 1.584,
+      "step": 2930
+    },
+    {
+      "epoch": 9.83,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9816069699903195,
+      "eval_loss": 0.016497749835252762,
+      "eval_precision": 0.9790344827586207,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.3849,
+      "eval_samples_per_second": 9.452,
+      "eval_steps_per_second": 1.586,
+      "step": 2940
+    },
+    {
+      "epoch": 9.87,
+      "eval_accuracy": 0.9971876575309756,
+      "eval_f1": 0.9821650767316467,
+      "eval_loss": 0.01616404764354229,
+      "eval_precision": 0.9793217535153019,
+      "eval_recall": 0.9850249584026622,
+      "eval_runtime": 45.4779,
+      "eval_samples_per_second": 9.433,
+      "eval_steps_per_second": 1.583,
+      "step": 2950
+    },
+    {
+      "epoch": 9.9,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9790829754813686,
+      "eval_loss": 0.016479264944791794,
+      "eval_precision": 0.9781345142540825,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.8218,
+      "eval_samples_per_second": 9.362,
+      "eval_steps_per_second": 1.571,
+      "step": 2960
+    },
+    {
+      "epoch": 9.93,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.980060924951537,
+      "eval_loss": 0.016018539667129517,
+      "eval_precision": 0.9787057522123894,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.4831,
+      "eval_samples_per_second": 9.432,
+      "eval_steps_per_second": 1.583,
+      "step": 2970
+    },
+    {
+      "epoch": 9.97,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9785348289710566,
+      "eval_loss": 0.016076229512691498,
+      "eval_precision": 0.9773167358229599,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.4108,
+      "eval_samples_per_second": 9.447,
+      "eval_steps_per_second": 1.586,
+      "step": 2980
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9797895902547065,
+      "eval_loss": 0.01618347130715847,
+      "eval_precision": 0.9781647318960752,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.4332,
+      "eval_samples_per_second": 9.442,
+      "eval_steps_per_second": 1.585,
+      "step": 2990
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 7e-06,
+      "loss": 0.0046,
+      "step": 3000
+    },
+    {
+      "epoch": 10.03,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9791234619106871,
+      "eval_loss": 0.016251519322395325,
+      "eval_precision": 0.9762889440308795,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.7254,
+      "eval_samples_per_second": 9.382,
+      "eval_steps_per_second": 1.575,
+      "step": 3000
+    },
+    {
+      "epoch": 10.07,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9806576402321082,
+      "eval_loss": 0.0159316323697567,
+      "eval_precision": 0.977147577092511,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.9937,
+      "eval_samples_per_second": 9.327,
+      "eval_steps_per_second": 1.565,
+      "step": 3010
+    },
+    {
+      "epoch": 10.1,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9798007747648035,
+      "eval_loss": 0.015915466472506523,
+      "eval_precision": 0.9776366648260629,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.3335,
+      "eval_samples_per_second": 9.463,
+      "eval_steps_per_second": 1.588,
+      "step": 3020
+    },
+    {
+      "epoch": 10.13,
+      "eval_accuracy": 0.9971478602318856,
+      "eval_f1": 0.9813200498132004,
+      "eval_loss": 0.015171729028224945,
+      "eval_precision": 0.9792874896437448,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.3108,
+      "eval_samples_per_second": 9.468,
+      "eval_steps_per_second": 1.589,
+      "step": 3030
+    },
+    {
+      "epoch": 10.17,
+      "eval_accuracy": 0.9966570268764426,
+      "eval_f1": 0.9667917187717104,
+      "eval_loss": 0.015076521784067154,
+      "eval_precision": 0.9688109161793372,
+      "eval_recall": 0.9647809206877427,
+      "eval_runtime": 45.3556,
+      "eval_samples_per_second": 9.459,
+      "eval_steps_per_second": 1.587,
+      "step": 3040
+    },
+    {
+      "epoch": 10.2,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.9746922970543493,
+      "eval_loss": 0.014860566705465317,
+      "eval_precision": 0.9721379310344828,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.4321,
+      "eval_samples_per_second": 9.443,
+      "eval_steps_per_second": 1.585,
+      "step": 3050
+    },
+    {
+      "epoch": 10.23,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.980110497237569,
+      "eval_loss": 0.014547467231750488,
+      "eval_precision": 0.9763346175013758,
+      "eval_recall": 0.9839156960621187,
+      "eval_runtime": 45.3097,
+      "eval_samples_per_second": 9.468,
+      "eval_steps_per_second": 1.589,
+      "step": 3060
+    },
+    {
+      "epoch": 10.27,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9792760431058304,
+      "eval_loss": 0.014211696572601795,
+      "eval_precision": 0.9757709251101322,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.7056,
+      "eval_samples_per_second": 9.386,
+      "eval_steps_per_second": 1.575,
+      "step": 3070
+    },
+    {
+      "epoch": 10.3,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9788469514724182,
+      "eval_loss": 0.01424412615597248,
+      "eval_precision": 0.9760132340777502,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.6738,
+      "eval_samples_per_second": 9.393,
+      "eval_steps_per_second": 1.576,
+      "step": 3080
+    },
+    {
+      "epoch": 10.33,
+      "eval_accuracy": 0.9969356079700724,
+      "eval_f1": 0.9745293466223699,
+      "eval_loss": 0.014736946672201157,
+      "eval_precision": 0.9729132117191819,
+      "eval_recall": 0.9761508596783139,
+      "eval_runtime": 45.4811,
+      "eval_samples_per_second": 9.432,
+      "eval_steps_per_second": 1.583,
+      "step": 3090
+    },
+    {
+      "epoch": 10.37,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9801050013815971,
+      "eval_loss": 0.014717744663357735,
+      "eval_precision": 0.9765969162995595,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.415,
+      "eval_samples_per_second": 9.446,
+      "eval_steps_per_second": 1.585,
+      "step": 3100
+    },
+    {
+      "epoch": 10.4,
+      "eval_accuracy": 0.9971080629327956,
+      "eval_f1": 0.9810642709053213,
+      "eval_loss": 0.015054759569466114,
+      "eval_precision": 0.9779553596031965,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.7255,
+      "eval_samples_per_second": 9.382,
+      "eval_steps_per_second": 1.575,
+      "step": 3110
+    },
+    {
+      "epoch": 10.43,
+      "eval_accuracy": 0.9971080629327956,
+      "eval_f1": 0.9807931463313527,
+      "eval_loss": 0.01575205847620964,
+      "eval_precision": 0.9774166896171853,
+      "eval_recall": 0.9841930116472546,
+      "eval_runtime": 45.3913,
+      "eval_samples_per_second": 9.451,
+      "eval_steps_per_second": 1.586,
+      "step": 3120
+    },
+    {
+      "epoch": 10.47,
+      "eval_accuracy": 0.9968162160728026,
+      "eval_f1": 0.97202216066482,
+      "eval_loss": 0.015902305021882057,
+      "eval_precision": 0.9709463198671832,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 45.3669,
+      "eval_samples_per_second": 9.456,
+      "eval_steps_per_second": 1.587,
+      "step": 3130
+    },
+    {
+      "epoch": 10.5,
+      "eval_accuracy": 0.9966835584091693,
+      "eval_f1": 0.9686633388796451,
+      "eval_loss": 0.016264360398054123,
+      "eval_precision": 0.9686633388796451,
+      "eval_recall": 0.9686633388796451,
+      "eval_runtime": 45.3009,
+      "eval_samples_per_second": 9.47,
+      "eval_steps_per_second": 1.589,
+      "step": 3140
+    },
+    {
+      "epoch": 10.54,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9782698961937716,
+      "eval_loss": 0.015213034115731716,
+      "eval_precision": 0.9765128488532744,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.7467,
+      "eval_samples_per_second": 9.378,
+      "eval_steps_per_second": 1.574,
+      "step": 3150
+    },
+    {
+      "epoch": 10.57,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.978693967902601,
+      "eval_loss": 0.014814727008342743,
+      "eval_precision": 0.9765323025952513,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.3386,
+      "eval_samples_per_second": 9.462,
+      "eval_steps_per_second": 1.588,
+      "step": 3160
+    },
+    {
+      "epoch": 10.6,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9793828697938286,
+      "eval_loss": 0.015096531249582767,
+      "eval_precision": 0.9773543220104943,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.6502,
+      "eval_samples_per_second": 9.398,
+      "eval_steps_per_second": 1.577,
+      "step": 3170
+    },
+    {
+      "epoch": 10.64,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9764998617638928,
+      "eval_loss": 0.015553129836916924,
+      "eval_precision": 0.9735391400220507,
+      "eval_recall": 0.9794786466999446,
+      "eval_runtime": 45.4708,
+      "eval_samples_per_second": 9.435,
+      "eval_steps_per_second": 1.583,
+      "step": 3180
+    },
+    {
+      "epoch": 10.67,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9774643992810729,
+      "eval_loss": 0.0158841609954834,
+      "eval_precision": 0.9746346843121036,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.4753,
+      "eval_samples_per_second": 9.434,
+      "eval_steps_per_second": 1.583,
+      "step": 3190
+    },
+    {
+      "epoch": 10.7,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9800884955752212,
+      "eval_loss": 0.016482515260577202,
+      "eval_precision": 0.9773855488141202,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.5435,
+      "eval_samples_per_second": 9.42,
+      "eval_steps_per_second": 1.581,
+      "step": 3200
+    },
+    {
+      "epoch": 10.74,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9795410561238596,
+      "eval_loss": 0.016413580626249313,
+      "eval_precision": 0.9765711135611908,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.616,
+      "eval_samples_per_second": 9.405,
+      "eval_steps_per_second": 1.578,
+      "step": 3210
+    },
+    {
+      "epoch": 10.77,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9791176877333702,
+      "eval_loss": 0.01622335985302925,
+      "eval_precision": 0.976551724137931,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.7676,
+      "eval_samples_per_second": 9.373,
+      "eval_steps_per_second": 1.573,
+      "step": 3220
+    },
+    {
+      "epoch": 10.8,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9803541781959049,
+      "eval_loss": 0.015803754329681396,
+      "eval_precision": 0.9781888459414688,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.7275,
+      "eval_samples_per_second": 9.382,
+      "eval_steps_per_second": 1.575,
+      "step": 3230
+    },
+    {
+      "epoch": 10.84,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9793828697938286,
+      "eval_loss": 0.015858009457588196,
+      "eval_precision": 0.9773543220104943,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.806,
+      "eval_samples_per_second": 9.366,
+      "eval_steps_per_second": 1.572,
+      "step": 3240
+    },
+    {
+      "epoch": 10.87,
+      "eval_accuracy": 0.9966968241755326,
+      "eval_f1": 0.9738263398421271,
+      "eval_loss": 0.016643954440951347,
+      "eval_precision": 0.9726141078838174,
+      "eval_recall": 0.9750415973377704,
+      "eval_runtime": 45.5359,
+      "eval_samples_per_second": 9.421,
+      "eval_steps_per_second": 1.581,
+      "step": 3250
+    },
+    {
+      "epoch": 10.9,
+      "eval_accuracy": 0.9969356079700724,
+      "eval_f1": 0.977728593166413,
+      "eval_loss": 0.016697365790605545,
+      "eval_precision": 0.9754347226055755,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.7975,
+      "eval_samples_per_second": 9.367,
+      "eval_steps_per_second": 1.572,
+      "step": 3260
+    },
+    {
+      "epoch": 10.94,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.977297895902547,
+      "eval_loss": 0.016878845170140266,
+      "eval_precision": 0.9756771697070205,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.4916,
+      "eval_samples_per_second": 9.43,
+      "eval_steps_per_second": 1.583,
+      "step": 3270
+    },
+    {
+      "epoch": 10.97,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.9780052566053397,
+      "eval_loss": 0.016803227365016937,
+      "eval_precision": 0.9757107369583218,
+      "eval_recall": 0.9803105934553522,
+      "eval_runtime": 45.574,
+      "eval_samples_per_second": 9.413,
+      "eval_steps_per_second": 1.58,
+      "step": 3280
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.9969754052691624,
+      "eval_f1": 0.9811790755604761,
+      "eval_loss": 0.016553932800889015,
+      "eval_precision": 0.9792817679558011,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.6533,
+      "eval_samples_per_second": 9.397,
+      "eval_steps_per_second": 1.577,
+      "step": 3290
+    },
+    {
+      "epoch": 11.04,
+      "eval_accuracy": 0.9971080629327956,
+      "eval_f1": 0.9845132743362832,
+      "eval_loss": 0.015565542504191399,
+      "eval_precision": 0.9817981246552675,
+      "eval_recall": 0.9872434830837493,
+      "eval_runtime": 45.6303,
+      "eval_samples_per_second": 9.402,
+      "eval_steps_per_second": 1.578,
+      "step": 3300
+    },
+    {
+      "epoch": 11.07,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9846579129232895,
+      "eval_loss": 0.014836783520877361,
+      "eval_precision": 0.981537613667677,
+      "eval_recall": 0.9877981142540211,
+      "eval_runtime": 45.5788,
+      "eval_samples_per_second": 9.412,
+      "eval_steps_per_second": 1.58,
+      "step": 3310
+    },
+    {
+      "epoch": 11.1,
+      "eval_accuracy": 0.9970549998673424,
+      "eval_f1": 0.9835521769177609,
+      "eval_loss": 0.015223776921629906,
+      "eval_precision": 0.9804353816478368,
+      "eval_recall": 0.9866888519134775,
+      "eval_runtime": 45.8006,
+      "eval_samples_per_second": 9.367,
+      "eval_steps_per_second": 1.572,
+      "step": 3320
+    },
+    {
+      "epoch": 11.14,
+      "eval_accuracy": 0.9969754052691624,
+      "eval_f1": 0.9827133176600746,
+      "eval_loss": 0.015493294224143028,
+      "eval_precision": 0.9801379310344828,
+      "eval_recall": 0.9853022739877981,
+      "eval_runtime": 46.0713,
+      "eval_samples_per_second": 9.312,
+      "eval_steps_per_second": 1.563,
+      "step": 3330
+    },
+    {
+      "epoch": 11.17,
+      "eval_accuracy": 0.9970019368018891,
+      "eval_f1": 0.9827180976081846,
+      "eval_loss": 0.014980579726397991,
+      "eval_precision": 0.9798731734215606,
+      "eval_recall": 0.985579589572934,
+      "eval_runtime": 45.5835,
+      "eval_samples_per_second": 9.411,
+      "eval_steps_per_second": 1.58,
+      "step": 3340
+    },
+    {
+      "epoch": 11.2,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9840918522617236,
+      "eval_loss": 0.014826910570263863,
+      "eval_precision": 0.9817830527187413,
+      "eval_recall": 0.9864115363283417,
+      "eval_runtime": 45.5689,
+      "eval_samples_per_second": 9.414,
+      "eval_steps_per_second": 1.58,
+      "step": 3350
+    },
+    {
+      "epoch": 11.24,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.983402489626556,
+      "eval_loss": 0.015205192379653454,
+      "eval_precision": 0.9809602649006622,
+      "eval_recall": 0.9858569051580699,
+      "eval_runtime": 45.5733,
+      "eval_samples_per_second": 9.413,
+      "eval_steps_per_second": 1.58,
+      "step": 3360
+    },
+    {
+      "epoch": 11.27,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9835430784123911,
+      "eval_loss": 0.015277662314474583,
+      "eval_precision": 0.9809655172413793,
+      "eval_recall": 0.9861342207432058,
+      "eval_runtime": 45.8383,
+      "eval_samples_per_second": 9.359,
+      "eval_steps_per_second": 1.571,
+      "step": 3370
+    },
+    {
+      "epoch": 11.3,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9803921568627452,
+      "eval_loss": 0.015845883637666702,
+      "eval_precision": 0.9763476347634763,
+      "eval_recall": 0.9844703272323905,
+      "eval_runtime": 45.6768,
+      "eval_samples_per_second": 9.392,
+      "eval_steps_per_second": 1.576,
+      "step": 3380
+    },
+    {
+      "epoch": 11.34,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9808037563872393,
+      "eval_loss": 0.016353582963347435,
+      "eval_precision": 0.9768913342503439,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.6996,
+      "eval_samples_per_second": 9.387,
+      "eval_steps_per_second": 1.576,
+      "step": 3390
+    },
+    {
+      "epoch": 11.37,
+      "eval_accuracy": 0.9966835584091693,
+      "eval_f1": 0.9728832318760378,
+      "eval_loss": 0.01731358841061592,
+      "eval_precision": 0.9707344008834898,
+      "eval_recall": 0.9750415973377704,
+      "eval_runtime": 45.9308,
+      "eval_samples_per_second": 9.34,
+      "eval_steps_per_second": 1.568,
+      "step": 3400
+    },
+    {
+      "epoch": 11.4,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9787116394802323,
+      "eval_loss": 0.015837378799915314,
+      "eval_precision": 0.9757442116868799,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.8585,
+      "eval_samples_per_second": 9.355,
+      "eval_steps_per_second": 1.57,
+      "step": 3410
+    },
+    {
+      "epoch": 11.44,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9741534208707672,
+      "eval_loss": 0.0164649561047554,
+      "eval_precision": 0.9710664094791954,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 45.5679,
+      "eval_samples_per_second": 9.415,
+      "eval_steps_per_second": 1.58,
+      "step": 3420
+    },
+    {
+      "epoch": 11.47,
+      "eval_accuracy": 0.9964182430819029,
+      "eval_f1": 0.9646373595895161,
+      "eval_loss": 0.017043212428689003,
+      "eval_precision": 0.9647711511789182,
+      "eval_recall": 0.9645036051026068,
+      "eval_runtime": 45.6723,
+      "eval_samples_per_second": 9.393,
+      "eval_steps_per_second": 1.576,
+      "step": 3430
+    },
+    {
+      "epoch": 11.51,
+      "eval_accuracy": 0.9964182430819029,
+      "eval_f1": 0.9653643668606262,
+      "eval_loss": 0.017299339175224304,
+      "eval_precision": 0.964562569213732,
+      "eval_recall": 0.9661674986134221,
+      "eval_runtime": 46.0069,
+      "eval_samples_per_second": 9.325,
+      "eval_steps_per_second": 1.565,
+      "step": 3440
+    },
+    {
+      "epoch": 11.54,
+      "eval_accuracy": 0.9964713061473561,
+      "eval_f1": 0.968054211035818,
+      "eval_loss": 0.01724633388221264,
+      "eval_precision": 0.9655172413793104,
+      "eval_recall": 0.9706045479755963,
+      "eval_runtime": 45.6115,
+      "eval_samples_per_second": 9.406,
+      "eval_steps_per_second": 1.579,
+      "step": 3450
+    },
+    {
+      "epoch": 11.57,
+      "eval_accuracy": 0.9964845719137194,
+      "eval_f1": 0.9670724958494743,
+      "eval_loss": 0.017298314720392227,
+      "eval_precision": 0.9649364991717283,
+      "eval_recall": 0.9692179700499168,
+      "eval_runtime": 45.5306,
+      "eval_samples_per_second": 9.422,
+      "eval_steps_per_second": 1.581,
+      "step": 3460
+    },
+    {
+      "epoch": 11.61,
+      "eval_accuracy": 0.9967366214746226,
+      "eval_f1": 0.9718875502008032,
+      "eval_loss": 0.015642836689949036,
+      "eval_precision": 0.9706777316735823,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 46.9214,
+      "eval_samples_per_second": 9.143,
+      "eval_steps_per_second": 1.534,
+      "step": 3470
+    },
+    {
+      "epoch": 11.64,
+      "eval_accuracy": 0.9969356079700724,
+      "eval_f1": 0.9768986028496335,
+      "eval_loss": 0.015028764493763447,
+      "eval_precision": 0.9746066795473365,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 45.682,
+      "eval_samples_per_second": 9.391,
+      "eval_steps_per_second": 1.576,
+      "step": 3480
+    },
+    {
+      "epoch": 11.67,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9775747508305648,
+      "eval_loss": 0.015305536799132824,
+      "eval_precision": 0.9759535655058044,
+      "eval_recall": 0.9792013311148087,
+      "eval_runtime": 45.8278,
+      "eval_samples_per_second": 9.361,
+      "eval_steps_per_second": 1.571,
+      "step": 3490
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 0.0042,
+      "step": 3500
+    },
+    {
+      "epoch": 11.71,
+      "eval_accuracy": 0.9969754052691624,
+      "eval_f1": 0.9772853185595569,
+      "eval_loss": 0.015260215848684311,
+      "eval_precision": 0.9762036524626453,
+      "eval_recall": 0.978369384359401,
+      "eval_runtime": 45.5813,
+      "eval_samples_per_second": 9.412,
+      "eval_steps_per_second": 1.58,
+      "step": 3500
+    },
+    {
+      "epoch": 11.74,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.9785407725321887,
+      "eval_loss": 0.015263444744050503,
+      "eval_precision": 0.9770528061929776,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 46.1625,
+      "eval_samples_per_second": 9.293,
+      "eval_steps_per_second": 1.56,
+      "step": 3510
+    },
+    {
+      "epoch": 11.77,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9793771626297577,
+      "eval_loss": 0.01575734093785286,
+      "eval_precision": 0.9776181265542968,
+      "eval_recall": 0.9811425402107599,
+      "eval_runtime": 45.6916,
+      "eval_samples_per_second": 9.389,
+      "eval_steps_per_second": 1.576,
+      "step": 3520
+    },
+    {
+      "epoch": 11.81,
+      "eval_accuracy": 0.9970682656337057,
+      "eval_f1": 0.9803541781959049,
+      "eval_loss": 0.016223162412643433,
+      "eval_precision": 0.9781888459414688,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.6681,
+      "eval_samples_per_second": 9.394,
+      "eval_steps_per_second": 1.577,
+      "step": 3530
+    },
+    {
+      "epoch": 11.84,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9794056668970283,
+      "eval_loss": 0.016862712800502777,
+      "eval_precision": 0.9763020115734362,
+      "eval_recall": 0.9825291181364393,
+      "eval_runtime": 45.8595,
+      "eval_samples_per_second": 9.355,
+      "eval_steps_per_second": 1.57,
+      "step": 3540
+    },
+    {
+      "epoch": 11.87,
+      "eval_accuracy": 0.9969754052691624,
+      "eval_f1": 0.9781526548672567,
+      "eval_loss": 0.016722770407795906,
+      "eval_precision": 0.9754550468836183,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.6514,
+      "eval_samples_per_second": 9.397,
+      "eval_steps_per_second": 1.577,
+      "step": 3550
+    },
+    {
+      "epoch": 11.91,
+      "eval_accuracy": 0.997041734100979,
+      "eval_f1": 0.9792473713337022,
+      "eval_loss": 0.01629817485809326,
+      "eval_precision": 0.9770844837106571,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.6901,
+      "eval_samples_per_second": 9.389,
+      "eval_steps_per_second": 1.576,
+      "step": 3560
+    },
+    {
+      "epoch": 11.94,
+      "eval_accuracy": 0.9970152025682524,
+      "eval_f1": 0.9795240730492528,
+      "eval_loss": 0.01616574078798294,
+      "eval_precision": 0.97736057426836,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 46.0985,
+      "eval_samples_per_second": 9.306,
+      "eval_steps_per_second": 1.562,
+      "step": 3570
+    },
+    {
+      "epoch": 11.97,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9778761061946902,
+      "eval_loss": 0.01619311049580574,
+      "eval_precision": 0.9751792608935466,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.9524,
+      "eval_samples_per_second": 9.336,
+      "eval_steps_per_second": 1.567,
+      "step": 3580
+    },
+    {
+      "epoch": 12.01,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9785822854774078,
+      "eval_loss": 0.01636885292828083,
+      "eval_precision": 0.97521343982374,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.6302,
+      "eval_samples_per_second": 9.402,
+      "eval_steps_per_second": 1.578,
+      "step": 3590
+    },
+    {
+      "epoch": 12.04,
+      "eval_accuracy": 0.9969886710355257,
+      "eval_f1": 0.9784470848300635,
+      "eval_loss": 0.01634366624057293,
+      "eval_precision": 0.9749449339207048,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.6457,
+      "eval_samples_per_second": 9.398,
+      "eval_steps_per_second": 1.577,
+      "step": 3600
+    },
+    {
+      "epoch": 12.07,
+      "eval_accuracy": 0.9969621395027991,
+      "eval_f1": 0.9776119402985075,
+      "eval_loss": 0.016408545896410942,
+      "eval_precision": 0.9743801652892562,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 46.0467,
+      "eval_samples_per_second": 9.317,
+      "eval_steps_per_second": 1.564,
+      "step": 3610
+    },
+    {
+      "epoch": 12.11,
+      "eval_accuracy": 0.9971345944655223,
+      "eval_f1": 0.9778822228366049,
+      "eval_loss": 0.01554455328732729,
+      "eval_precision": 0.9749173098125689,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 45.8672,
+      "eval_samples_per_second": 9.353,
+      "eval_steps_per_second": 1.57,
+      "step": 3620
+    },
+    {
+      "epoch": 12.14,
+      "eval_accuracy": 0.9971743917646122,
+      "eval_f1": 0.9781284606866002,
+      "eval_loss": 0.015367398969829082,
+      "eval_precision": 0.9765063571033721,
+      "eval_recall": 0.9797559622850804,
+      "eval_runtime": 45.8306,
+      "eval_samples_per_second": 9.361,
+      "eval_steps_per_second": 1.571,
+      "step": 3630
+    },
+    {
+      "epoch": 12.17,
+      "eval_accuracy": 0.9970815314000689,
+      "eval_f1": 0.9784172661870504,
+      "eval_loss": 0.016077237203717232,
+      "eval_precision": 0.9762562120375483,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.7148,
+      "eval_samples_per_second": 9.384,
+      "eval_steps_per_second": 1.575,
+      "step": 3640
+    },
+    {
+      "epoch": 12.21,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9757986447241046,
+      "eval_loss": 0.016824740916490555,
+      "eval_precision": 0.9732413793103448,
+      "eval_recall": 0.978369384359401,
+      "eval_runtime": 46.1214,
+      "eval_samples_per_second": 9.302,
+      "eval_steps_per_second": 1.561,
+      "step": 3650
+    },
+    {
+      "epoch": 12.24,
+      "eval_accuracy": 0.9969223422037091,
+      "eval_f1": 0.978146611341632,
+      "eval_loss": 0.016822459176182747,
+      "eval_precision": 0.9757174392935982,
+      "eval_recall": 0.9805879090404881,
+      "eval_runtime": 45.7927,
+      "eval_samples_per_second": 9.368,
+      "eval_steps_per_second": 1.572,
+      "step": 3660
+    },
+    {
+      "epoch": 12.27,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9792588495575221,
+      "eval_loss": 0.01815040595829487,
+      "eval_precision": 0.9765581908439052,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.8912,
+      "eval_samples_per_second": 9.348,
+      "eval_steps_per_second": 1.569,
+      "step": 3670
+    },
+    {
+      "epoch": 12.31,
+      "eval_accuracy": 0.9967764187737126,
+      "eval_f1": 0.9788586430841509,
+      "eval_loss": 0.01881832629442215,
+      "eval_precision": 0.9754888460479206,
+      "eval_recall": 0.9822518025513034,
+      "eval_runtime": 46.0893,
+      "eval_samples_per_second": 9.308,
+      "eval_steps_per_second": 1.562,
+      "step": 3680
+    },
+    {
+      "epoch": 12.34,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.976892209768922,
+      "eval_loss": 0.017492154613137245,
+      "eval_precision": 0.9748688207677437,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.8024,
+      "eval_samples_per_second": 9.366,
+      "eval_steps_per_second": 1.572,
+      "step": 3690
+    },
+    {
+      "epoch": 12.37,
+      "eval_accuracy": 0.9966172295773527,
+      "eval_f1": 0.9710806697108068,
+      "eval_loss": 0.018154501914978027,
+      "eval_precision": 0.9690693178679922,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 46.0834,
+      "eval_samples_per_second": 9.309,
+      "eval_steps_per_second": 1.562,
+      "step": 3700
+    },
+    {
+      "epoch": 12.41,
+      "eval_accuracy": 0.9962457881191796,
+      "eval_f1": 0.9614745011086475,
+      "eval_loss": 0.01924285851418972,
+      "eval_precision": 0.9609418282548476,
+      "eval_recall": 0.9620077648363838,
+      "eval_runtime": 45.696,
+      "eval_samples_per_second": 9.388,
+      "eval_steps_per_second": 1.576,
+      "step": 3710
+    },
+    {
+      "epoch": 12.44,
+      "eval_accuracy": 0.9964580403809928,
+      "eval_f1": 0.9651259341267644,
+      "eval_loss": 0.017841314896941185,
+      "eval_precision": 0.9632596685082873,
+      "eval_recall": 0.9669994453688298,
+      "eval_runtime": 45.9939,
+      "eval_samples_per_second": 9.327,
+      "eval_steps_per_second": 1.565,
+      "step": 3720
+    },
+    {
+      "epoch": 12.47,
+      "eval_accuracy": 0.9966304953437161,
+      "eval_f1": 0.970962389380531,
+      "eval_loss": 0.016754455864429474,
+      "eval_precision": 0.968284611141754,
+      "eval_recall": 0.973655019412091,
+      "eval_runtime": 45.8171,
+      "eval_samples_per_second": 9.363,
+      "eval_steps_per_second": 1.571,
+      "step": 3730
+    },
+    {
+      "epoch": 12.51,
+      "eval_accuracy": 0.9968294818391659,
+      "eval_f1": 0.9774269491760144,
+      "eval_loss": 0.015923812985420227,
+      "eval_precision": 0.9762102351313969,
+      "eval_recall": 0.9786466999445369,
+      "eval_runtime": 45.7801,
+      "eval_samples_per_second": 9.371,
+      "eval_steps_per_second": 1.573,
+      "step": 3740
+    },
+    {
+      "epoch": 12.54,
+      "eval_accuracy": 0.9968825449046191,
+      "eval_f1": 0.9772916089725838,
+      "eval_loss": 0.015936698764562607,
+      "eval_precision": 0.9759402654867256,
+      "eval_recall": 0.9786466999445369,
+      "eval_runtime": 45.9975,
+      "eval_samples_per_second": 9.327,
+      "eval_steps_per_second": 1.565,
+      "step": 3750
+    },
+    {
+      "epoch": 12.58,
+      "eval_accuracy": 0.9968029503064392,
+      "eval_f1": 0.9736623232603271,
+      "eval_loss": 0.016157541424036026,
+      "eval_precision": 0.9733924611973392,
+      "eval_recall": 0.9739323349972269,
+      "eval_runtime": 45.7117,
+      "eval_samples_per_second": 9.385,
+      "eval_steps_per_second": 1.575,
+      "step": 3760
+    },
+    {
+      "epoch": 12.61,
+      "eval_accuracy": 0.9967366214746226,
+      "eval_f1": 0.9694117647058823,
+      "eval_loss": 0.01580197736620903,
+      "eval_precision": 0.9676706272450953,
+      "eval_recall": 0.971159179145868,
+      "eval_runtime": 45.5828,
+      "eval_samples_per_second": 9.411,
+      "eval_steps_per_second": 1.58,
+      "step": 3770
+    },
+    {
+      "epoch": 12.64,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.971460238293156,
+      "eval_loss": 0.015100532211363316,
+      "eval_precision": 0.9706533776301218,
+      "eval_recall": 0.9722684414864116,
+      "eval_runtime": 45.7142,
+      "eval_samples_per_second": 9.384,
+      "eval_steps_per_second": 1.575,
+      "step": 3780
+    },
+    {
+      "epoch": 12.68,
+      "eval_accuracy": 0.9969488737364357,
+      "eval_f1": 0.9725609756097561,
+      "eval_loss": 0.015240387991070747,
+      "eval_precision": 0.97202216066482,
+      "eval_recall": 0.9731003882418192,
+      "eval_runtime": 46.103,
+      "eval_samples_per_second": 9.305,
+      "eval_steps_per_second": 1.562,
+      "step": 3790
+    },
+    {
+      "epoch": 12.71,
+      "eval_accuracy": 0.9970284683346157,
+      "eval_f1": 0.9779991697799916,
+      "eval_loss": 0.015490233898162842,
+      "eval_precision": 0.975973487986744,
+      "eval_recall": 0.9800332778702163,
+      "eval_runtime": 45.7054,
+      "eval_samples_per_second": 9.386,
+      "eval_steps_per_second": 1.575,
+      "step": 3800
+    },
+    {
+      "epoch": 12.74,
+      "eval_accuracy": 0.9971213286991589,
+      "eval_f1": 0.980060924951537,
+      "eval_loss": 0.014931376092135906,
+      "eval_precision": 0.9787057522123894,
+      "eval_recall": 0.9814198557958957,
+      "eval_runtime": 45.8361,
+      "eval_samples_per_second": 9.359,
+      "eval_steps_per_second": 1.571,
+      "step": 3810
+    },
+    {
+      "epoch": 12.78,
+      "eval_accuracy": 0.9970947971664323,
+      "eval_f1": 0.9769764216366158,
+      "eval_loss": 0.013774119317531586,
+      "eval_precision": 0.9772475027746947,
+      "eval_recall": 0.9767054908485857,
+      "eval_runtime": 46.4293,
+      "eval_samples_per_second": 9.24,
+      "eval_steps_per_second": 1.551,
+      "step": 3820
+    },
+    {
+      "epoch": 12.81,
+      "eval_accuracy": 0.9971611259982489,
+      "eval_f1": 0.9775436650956474,
+      "eval_loss": 0.01355548482388258,
+      "eval_precision": 0.9772727272727273,
+      "eval_recall": 0.9778147531891292,
+      "eval_runtime": 45.7415,
+      "eval_samples_per_second": 9.379,
+      "eval_steps_per_second": 1.574,
+      "step": 3830
+    },
+    {
+      "epoch": 12.84,
+      "eval_accuracy": 0.9973335809609721,
+      "eval_f1": 0.9824124082537044,
+      "eval_loss": 0.013703616335988045,
+      "eval_precision": 0.9811894882434301,
+      "eval_recall": 0.9836383804769828,
+      "eval_runtime": 45.73,
+      "eval_samples_per_second": 9.381,
+      "eval_steps_per_second": 1.574,
+      "step": 3840
+    },
+    {
+      "epoch": 12.88,
+      "eval_accuracy": 0.9973601124936987,
+      "eval_f1": 0.9812102790826194,
+      "eval_loss": 0.014531377702951431,
+      "eval_precision": 0.9776982378854625,
+      "eval_recall": 0.9847476428175264,
+      "eval_runtime": 45.7852,
+      "eval_samples_per_second": 9.37,
+      "eval_steps_per_second": 1.573,
+      "step": 3850
+    },
+    {
+      "epoch": 12.91,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.977526540741762,
+      "eval_loss": 0.016427284106612206,
+      "eval_precision": 0.9720318069646284,
+      "eval_recall": 0.983083749306711,
+      "eval_runtime": 45.8438,
+      "eval_samples_per_second": 9.358,
+      "eval_steps_per_second": 1.571,
+      "step": 3860
+    },
+    {
+      "epoch": 12.94,
+      "eval_accuracy": 0.9960202700910031,
+      "eval_f1": 0.9653931339977851,
+      "eval_loss": 0.021676059812307358,
+      "eval_precision": 0.9637921503593145,
+      "eval_recall": 0.9669994453688298,
+      "eval_runtime": 45.7888,
+      "eval_samples_per_second": 9.369,
+      "eval_steps_per_second": 1.572,
+      "step": 3870
+    },
+    {
+      "epoch": 12.98,
+      "eval_accuracy": 0.9967498872409859,
+      "eval_f1": 0.9777900400055181,
+      "eval_loss": 0.018003830686211586,
+      "eval_precision": 0.9728245951139171,
+      "eval_recall": 0.9828064337215752,
+      "eval_runtime": 45.7325,
+      "eval_samples_per_second": 9.381,
+      "eval_steps_per_second": 1.574,
+      "step": 3880
+    },
+    {
+      "epoch": 13.01,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.977667493796526,
+      "eval_loss": 0.01754719577729702,
+      "eval_precision": 0.9720394736842105,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 46.1857,
+      "eval_samples_per_second": 9.289,
+      "eval_steps_per_second": 1.559,
+      "step": 3890
+    },
+    {
+      "epoch": 13.04,
+      "eval_accuracy": 0.9967631530073492,
+      "eval_f1": 0.9739274382673472,
+      "eval_loss": 0.01792616955935955,
+      "eval_precision": 0.96898160856437,
+      "eval_recall": 0.9789240155296728,
+      "eval_runtime": 45.7326,
+      "eval_samples_per_second": 9.381,
+      "eval_steps_per_second": 1.574,
+      "step": 3900
+    },
+    {
+      "epoch": 13.08,
+      "eval_accuracy": 0.9967631530073492,
+      "eval_f1": 0.9729431253451133,
+      "eval_loss": 0.017918290570378304,
+      "eval_precision": 0.9686641011544805,
+      "eval_recall": 0.9772601220188575,
+      "eval_runtime": 46.0978,
+      "eval_samples_per_second": 9.306,
+      "eval_steps_per_second": 1.562,
+      "step": 3910
+    },
+    {
+      "epoch": 13.11,
+      "eval_accuracy": 0.9968692791382558,
+      "eval_f1": 0.9765323025952513,
+      "eval_loss": 0.017354754731059074,
+      "eval_precision": 0.9722374931280924,
+      "eval_recall": 0.980865224625624,
+      "eval_runtime": 46.21,
+      "eval_samples_per_second": 9.284,
+      "eval_steps_per_second": 1.558,
+      "step": 3920
+    },
+    {
+      "epoch": 13.14,
+      "eval_accuracy": 0.9969356079700724,
+      "eval_f1": 0.9775017253278124,
+      "eval_loss": 0.015932898968458176,
+      "eval_precision": 0.9730695245946689,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 45.7375,
+      "eval_samples_per_second": 9.38,
+      "eval_steps_per_second": 1.574,
+      "step": 3930
+    },
+    {
+      "epoch": 13.18,
+      "eval_accuracy": 0.9969090764373458,
+      "eval_f1": 0.9779005524861878,
+      "eval_loss": 0.015960365533828735,
+      "eval_precision": 0.9741331865712713,
+      "eval_recall": 0.9816971713810316,
+      "eval_runtime": 45.8882,
+      "eval_samples_per_second": 9.349,
+      "eval_steps_per_second": 1.569,
+      "step": 3940
+    },
+    {
+      "epoch": 13.21,
+      "eval_accuracy": 0.9967498872409859,
+      "eval_f1": 0.9737569060773481,
+      "eval_loss": 0.017051612958312035,
+      "eval_precision": 0.9700055035773253,
+      "eval_recall": 0.9775374376039934,
+      "eval_runtime": 45.8012,
+      "eval_samples_per_second": 9.367,
+      "eval_steps_per_second": 1.572,
+      "step": 3950
+    },
+    {
+      "epoch": 13.24,
+      "eval_accuracy": 0.9966702926428059,
+      "eval_f1": 0.9723527785457561,
+      "eval_loss": 0.01766437292098999,
+      "eval_precision": 0.9694046306504961,
+      "eval_recall": 0.9753189129229063,
+      "eval_runtime": 45.8759,
+      "eval_samples_per_second": 9.351,
+      "eval_steps_per_second": 1.569,
+      "step": 3960
+    },
+    {
+      "epoch": 13.28,
+      "eval_accuracy": 0.9966570268764426,
+      "eval_f1": 0.9727411097274111,
+      "eval_loss": 0.01822058856487274,
+      "eval_precision": 0.9707263186964927,
+      "eval_recall": 0.9747642817526345,
+      "eval_runtime": 45.8529,
+      "eval_samples_per_second": 9.356,
+      "eval_steps_per_second": 1.57,
+      "step": 3970
+    },
+    {
+      "epoch": 13.31,
+      "eval_accuracy": 0.9966304953437161,
+      "eval_f1": 0.9724872113922302,
+      "eval_loss": 0.01849460043013096,
+      "eval_precision": 0.9696719051557762,
+      "eval_recall": 0.9753189129229063,
+      "eval_runtime": 45.7611,
+      "eval_samples_per_second": 9.375,
+      "eval_steps_per_second": 1.573,
+      "step": 3980
+    },
+    {
+      "epoch": 13.34,
+      "eval_accuracy": 0.9968427476055292,
+      "eval_f1": 0.9780417069465543,
+      "eval_loss": 0.017539281398057938,
+      "eval_precision": 0.9741403026134801,
+      "eval_recall": 0.9819744869661675,
+      "eval_runtime": 46.0339,
+      "eval_samples_per_second": 9.319,
+      "eval_steps_per_second": 1.564,
+      "step": 3990
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 6e-06,
+      "loss": 0.0037,
+      "step": 4000
+    },
+    {
+      "epoch": 13.38,
+      "eval_accuracy": 0.9968958106709824,
+      "eval_f1": 0.9792874896437448,
+      "eval_loss": 0.017371075227856636,
+      "eval_precision": 0.9752475247524752,
+      "eval_recall": 0.9833610648918469,
+      "eval_runtime": 45.7746,
+      "eval_samples_per_second": 9.372,
+      "eval_steps_per_second": 1.573,
+      "step": 4000
+    }
+  ],
+  "max_steps": 10000,
+  "num_train_epochs": 34,
+  "total_flos": 6370581676032000.0,
+  "trial_name": null,
+  "trial_params": null
+}