diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -1,2518 +1,10806 @@
 {
-  "best_global_step": 800,
-  "best_metric": 74.85714285714286,
-  "best_model_checkpoint": "/data/datht163/benchmark/checkpoint-800",
-  "epoch": 19.752351097178682,
+  "best_global_step": 3000,
+  "best_metric": 88.02395209580838,
+  "best_model_checkpoint": "/data/datht163/benchmark/checkpoint-3000",
+  "epoch": 19.998003992015967,
   "eval_steps": 500,
-  "global_step": 1580,
+  "global_step": 7500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.06269592476489028,
-      "grad_norm": 20.461719512939453,
-      "learning_rate": 6.25e-05,
-      "loss": 4.5641,
+      "epoch": 0.01330671989354624,
+      "grad_norm": 60.81864547729492,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 4.6562,
       "step": 5
     },
     {
-      "epoch": 0.12539184952978055,
-      "grad_norm": 45.15224075317383,
-      "learning_rate": 0.000125,
-      "loss": 4.8469,
+      "epoch": 0.02661343978709248,
+      "grad_norm": 37.34549331665039,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 4.0133,
       "step": 10
     },
     {
-      "epoch": 0.18808777429467086,
-      "grad_norm": 22.794708251953125,
-      "learning_rate": 0.0001875,
-      "loss": 4.368,
+      "epoch": 0.03992015968063872,
+      "grad_norm": 11.796462059020996,
+      "learning_rate": 4e-05,
+      "loss": 2.6719,
       "step": 15
     },
     {
-      "epoch": 0.2507836990595611,
-      "grad_norm": 19.09231185913086,
-      "learning_rate": 0.00019999677214588312,
-      "loss": 3.9258,
+      "epoch": 0.05322687957418496,
+      "grad_norm": 8.3694429397583,
+      "learning_rate": 5.333333333333333e-05,
+      "loss": 3.0212,
       "step": 20
     },
     {
-      "epoch": 0.31347962382445144,
-      "grad_norm": 15.92790412902832,
-      "learning_rate": 0.0001999836593456696,
-      "loss": 3.2234,
+      "epoch": 0.0665335994677312,
+      "grad_norm": 15.191309928894043,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 2.2947,
       "step": 25
     },
     {
-      "epoch": 0.3761755485893417,
-      "grad_norm": 19.797821044921875,
-      "learning_rate": 0.00019996046118014955,
-      "loss": 3.2684,
+      "epoch": 0.07984031936127745,
+      "grad_norm": 5.090390682220459,
+      "learning_rate": 8e-05,
+      "loss": 1.9277,
       "step": 30
     },
     {
-      "epoch": 0.438871473354232,
-      "grad_norm": 19.880847930908203,
-      "learning_rate": 0.00019992717998932507,
-      "loss": 3.1656,
+      "epoch": 0.09314703925482369,
+      "grad_norm": 11.129491806030273,
+      "learning_rate": 9.333333333333334e-05,
+      "loss": 2.1779,
       "step": 35
     },
     {
-      "epoch": 0.5015673981191222,
-      "grad_norm": 19.707406997680664,
-      "learning_rate": 0.00019988381913027442,
-      "loss": 3.0031,
+      "epoch": 0.10645375914836992,
+      "grad_norm": 7.373464584350586,
+      "learning_rate": 0.00010666666666666667,
+      "loss": 2.0246,
       "step": 40
     },
     {
-      "epoch": 0.5642633228840125,
-      "grad_norm": 14.307902336120605,
-      "learning_rate": 0.00019983038297681336,
-      "loss": 2.7219,
+      "epoch": 0.11976047904191617,
+      "grad_norm": 11.810800552368164,
+      "learning_rate": 0.00012,
+      "loss": 1.8453,
       "step": 45
     },
     {
-      "epoch": 0.6269592476489029,
-      "grad_norm": 13.697257995605469,
-      "learning_rate": 0.00019976687691905393,
-      "loss": 2.6117,
+      "epoch": 0.1330671989354624,
+      "grad_norm": 11.086739540100098,
+      "learning_rate": 0.00013333333333333334,
+      "loss": 2.0291,
       "step": 50
     },
     {
-      "epoch": 0.6896551724137931,
-      "grad_norm": 20.15327262878418,
-      "learning_rate": 0.0001996933073628608,
-      "loss": 2.7996,
+      "epoch": 0.14637391882900866,
+      "grad_norm": 9.655720710754395,
+      "learning_rate": 0.00014666666666666666,
+      "loss": 1.8613,
       "step": 55
     },
     {
-      "epoch": 0.7523510971786834,
-      "grad_norm": 20.71539878845215,
-      "learning_rate": 0.00019960968172920516,
-      "loss": 2.5578,
+      "epoch": 0.1596806387225549,
+      "grad_norm": 9.862675666809082,
+      "learning_rate": 0.00016,
+      "loss": 1.5939,
       "step": 60
     },
     {
-      "epoch": 0.8150470219435737,
-      "grad_norm": 8.937211036682129,
-      "learning_rate": 0.00019951600845341595,
-      "loss": 2.9141,
+      "epoch": 0.17298735861610112,
+      "grad_norm": 7.531759262084961,
+      "learning_rate": 0.00017333333333333334,
+      "loss": 2.0128,
       "step": 65
     },
     {
-      "epoch": 0.877742946708464,
-      "grad_norm": 8.478225708007812,
-      "learning_rate": 0.0001994122969843293,
-      "loss": 2.8832,
+      "epoch": 0.18629407850964738,
+      "grad_norm": 9.651808738708496,
+      "learning_rate": 0.0001866666666666667,
+      "loss": 1.7102,
       "step": 70
     },
     {
-      "epoch": 0.9404388714733543,
-      "grad_norm": 15.695404052734375,
-      "learning_rate": 0.00019929855778333516,
-      "loss": 2.7437,
+      "epoch": 0.1996007984031936,
+      "grad_norm": 9.359721183776855,
+      "learning_rate": 0.0002,
+      "loss": 1.7814,
       "step": 75
     },
     {
-      "epoch": 1.0,
-      "grad_norm": 19.672351837158203,
-      "learning_rate": 0.00019917480232332224,
-      "loss": 2.2395,
+      "epoch": 0.21290751829673984,
+      "grad_norm": 14.421661376953125,
+      "learning_rate": 0.00019999977622236462,
+      "loss": 1.8324,
       "step": 80
     },
     {
-      "epoch": 1.0,
-      "eval_loss": 0.6336904764175415,
-      "eval_macro_f1": 73.26745178358081,
-      "eval_macro_precision": 75.24521017228876,
-      "eval_macro_recall": 73.14285714285714,
-      "eval_micro_f1": 73.14285714285714,
-      "eval_micro_precision": 73.14285714285714,
-      "eval_micro_recall": 73.14285714285714,
-      "eval_runtime": 5.7289,
-      "eval_samples_per_second": 183.283,
-      "eval_steps_per_second": 11.521,
-      "step": 80
-    },
-    {
-      "epoch": 1.0626959247648902,
-      "grad_norm": 14.15684700012207,
-      "learning_rate": 0.0001990410430875205,
-      "loss": 1.4396,
+      "epoch": 0.2262142381902861,
+      "grad_norm": 8.235530853271484,
+      "learning_rate": 0.00019999910489045997,
+      "loss": 1.8022,
       "step": 85
     },
     {
-      "epoch": 1.1253918495297806,
-      "grad_norm": 20.00255012512207,
-      "learning_rate": 0.00019889729356824235,
-      "loss": 1.3728,
+      "epoch": 0.23952095808383234,
+      "grad_norm": 19.68333625793457,
+      "learning_rate": 0.00019999798600729064,
+      "loss": 1.6848,
       "step": 90
     },
     {
-      "epoch": 1.188087774294671,
-      "grad_norm": 13.56872272491455,
-      "learning_rate": 0.00019874356826552129,
-      "loss": 1.6022,
+      "epoch": 0.2528276779773786,
+      "grad_norm": 9.926045417785645,
+      "learning_rate": 0.00019999641957786432,
+      "loss": 1.4403,
       "step": 95
     },
     {
-      "epoch": 1.250783699059561,
-      "grad_norm": 6.213724613189697,
-      "learning_rate": 0.00019857988268564953,
-      "loss": 1.2819,
+      "epoch": 0.2661343978709248,
+      "grad_norm": 8.2724027633667,
+      "learning_rate": 0.00019999440560919152,
+      "loss": 2.1332,
       "step": 100
     },
     {
-      "epoch": 1.3134796238244515,
-      "grad_norm": 28.165010452270508,
-      "learning_rate": 0.0001984062533396137,
-      "loss": 1.1332,
+      "epoch": 0.27944111776447106,
+      "grad_norm": 13.359396934509277,
+      "learning_rate": 0.00019999194411028594,
+      "loss": 1.7551,
       "step": 105
     },
     {
-      "epoch": 1.3761755485893417,
-      "grad_norm": 15.799768447875977,
-      "learning_rate": 0.00019822269774142954,
-      "loss": 1.244,
+      "epoch": 0.2927478376580173,
+      "grad_norm": 11.357268333435059,
+      "learning_rate": 0.00019998903509216415,
+      "loss": 1.9558,
       "step": 110
     },
     {
-      "epoch": 1.438871473354232,
-      "grad_norm": 21.20913314819336,
-      "learning_rate": 0.0001980292344063752,
-      "loss": 1.4427,
+      "epoch": 0.3060545575515635,
+      "grad_norm": 13.34130859375,
+      "learning_rate": 0.0001999856785678456,
+      "loss": 1.6962,
       "step": 115
     },
     {
-      "epoch": 1.5015673981191222,
-      "grad_norm": 18.147506713867188,
-      "learning_rate": 0.0001978258828491236,
-      "loss": 1.4596,
+      "epoch": 0.3193612774451098,
+      "grad_norm": 4.8148698806762695,
+      "learning_rate": 0.0001999818745523526,
+      "loss": 1.6982,
       "step": 120
     },
     {
-      "epoch": 1.5642633228840124,
-      "grad_norm": 17.44733428955078,
-      "learning_rate": 0.00019761266358177398,
-      "loss": 0.9827,
+      "epoch": 0.33266799733865604,
+      "grad_norm": 6.624136447906494,
+      "learning_rate": 0.0001999776230627102,
+      "loss": 1.8629,
       "step": 125
     },
     {
-      "epoch": 1.626959247648903,
-      "grad_norm": 18.33002471923828,
-      "learning_rate": 0.00019738959811178272,
-      "loss": 1.2366,
+      "epoch": 0.34597471723220224,
+      "grad_norm": 6.114648818969727,
+      "learning_rate": 0.00019997292411794618,
+      "loss": 1.6615,
       "step": 130
     },
     {
-      "epoch": 1.6896551724137931,
-      "grad_norm": 16.217641830444336,
-      "learning_rate": 0.00019715670893979414,
-      "loss": 1.2665,
+      "epoch": 0.3592814371257485,
+      "grad_norm": 7.139102458953857,
+      "learning_rate": 0.00019996777773909093,
+      "loss": 1.4755,
       "step": 135
     },
     {
-      "epoch": 1.7523510971786833,
-      "grad_norm": 16.21353530883789,
-      "learning_rate": 0.00019691401955737072,
-      "loss": 1.5896,
+      "epoch": 0.37258815701929476,
+      "grad_norm": 15.46536636352539,
+      "learning_rate": 0.0001999621839491773,
+      "loss": 1.8553,
       "step": 140
     },
     {
-      "epoch": 1.8150470219435737,
-      "grad_norm": 11.688411712646484,
-      "learning_rate": 0.0001966615544446234,
-      "loss": 1.5775,
+      "epoch": 0.38589487691284097,
+      "grad_norm": 6.5360331535339355,
+      "learning_rate": 0.00019995614277324065,
+      "loss": 1.5998,
       "step": 145
     },
     {
-      "epoch": 1.877742946708464,
-      "grad_norm": 14.346232414245605,
-      "learning_rate": 0.0001963993390677424,
-      "loss": 1.1833,
+      "epoch": 0.3992015968063872,
+      "grad_norm": 8.404138565063477,
+      "learning_rate": 0.00019994965423831854,
+      "loss": 1.4908,
       "step": 150
     },
     {
-      "epoch": 1.9404388714733543,
-      "grad_norm": 6.273299694061279,
-      "learning_rate": 0.00019612739987642845,
-      "loss": 1.0552,
+      "epoch": 0.4125083166999335,
+      "grad_norm": 8.757245063781738,
+      "learning_rate": 0.00019994271837345072,
+      "loss": 1.6893,
       "step": 155
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 17.62116050720215,
-      "learning_rate": 0.00019584576430122473,
-      "loss": 1.0835,
-      "step": 160
-    },
-    {
-      "epoch": 2.0,
-      "eval_loss": 0.7825595140457153,
-      "eval_macro_f1": 73.56897234236968,
-      "eval_macro_precision": 73.7348812342735,
-      "eval_macro_recall": 73.90476190476191,
-      "eval_micro_f1": 73.90476190476191,
-      "eval_micro_precision": 73.90476190476191,
-      "eval_micro_recall": 73.90476190476191,
-      "eval_runtime": 3.4148,
-      "eval_samples_per_second": 307.482,
-      "eval_steps_per_second": 19.327,
+      "epoch": 0.4258150365934797,
+      "grad_norm": 5.324100017547607,
+      "learning_rate": 0.00019993533520967912,
+      "loss": 1.4273,
       "step": 160
     },
     {
-      "epoch": 2.06269592476489,
-      "grad_norm": 4.175657749176025,
-      "learning_rate": 0.00019555446075075,
-      "loss": 0.3124,
+      "epoch": 0.43912175648702595,
+      "grad_norm": 6.764120101928711,
+      "learning_rate": 0.00019992750478004738,
+      "loss": 1.3152,
       "step": 165
     },
     {
-      "epoch": 2.1253918495297803,
-      "grad_norm": 9.987613677978516,
-      "learning_rate": 0.00019525351860883293,
-      "loss": 0.3053,
+      "epoch": 0.4524284763805722,
+      "grad_norm": 16.72058868408203,
+      "learning_rate": 0.00019991922711960102,
+      "loss": 1.5234,
       "step": 170
     },
     {
-      "epoch": 2.188087774294671,
-      "grad_norm": 27.39342498779297,
-      "learning_rate": 0.00019494296823154835,
-      "loss": 0.2825,
+      "epoch": 0.4657351962741184,
+      "grad_norm": 7.501124858856201,
+      "learning_rate": 0.0001999105022653872,
+      "loss": 1.51,
       "step": 175
     },
     {
-      "epoch": 2.250783699059561,
-      "grad_norm": 5.328784942626953,
-      "learning_rate": 0.000194622840944155,
-      "loss": 0.32,
+      "epoch": 0.47904191616766467,
+      "grad_norm": 9.424518585205078,
+      "learning_rate": 0.0001999013302564544,
+      "loss": 1.6568,
       "step": 180
     },
     {
-      "epoch": 2.3134796238244513,
-      "grad_norm": 8.484028816223145,
-      "learning_rate": 0.00019429316903793583,
-      "loss": 0.0699,
+      "epoch": 0.49234863606121093,
+      "grad_norm": 5.219127655029297,
+      "learning_rate": 0.0001998917111338525,
+      "loss": 1.3971,
       "step": 185
     },
     {
-      "epoch": 2.376175548589342,
-      "grad_norm": 16.879404067993164,
-      "learning_rate": 0.00019395398576694086,
-      "loss": 0.2403,
+      "epoch": 0.5056553559547572,
+      "grad_norm": 8.969573020935059,
+      "learning_rate": 0.00019988164494063226,
+      "loss": 1.5215,
       "step": 190
     },
     {
-      "epoch": 2.438871473354232,
-      "grad_norm": 49.66862487792969,
-      "learning_rate": 0.0001936053253446327,
-      "loss": 0.2737,
+      "epoch": 0.5189620758483033,
+      "grad_norm": 9.649981498718262,
+      "learning_rate": 0.00019987113172184563,
+      "loss": 1.5256,
       "step": 195
     },
     {
-      "epoch": 2.501567398119122,
-      "grad_norm": 17.669599533081055,
-      "learning_rate": 0.00019324722294043558,
-      "loss": 0.4838,
+      "epoch": 0.5322687957418496,
+      "grad_norm": 7.442275047302246,
+      "learning_rate": 0.00019986017152454495,
+      "loss": 1.3934,
       "step": 200
     },
     {
-      "epoch": 2.5642633228840124,
-      "grad_norm": 14.710466384887695,
-      "learning_rate": 0.00019287971467618766,
-      "loss": 0.2973,
+      "epoch": 0.5455755156353959,
+      "grad_norm": 6.018413543701172,
+      "learning_rate": 0.0001998487643977832,
+      "loss": 1.4781,
       "step": 205
     },
     {
-      "epoch": 2.626959247648903,
-      "grad_norm": 25.36009979248047,
-      "learning_rate": 0.00019250283762249748,
-      "loss": 0.4224,
+      "epoch": 0.5588822355289421,
+      "grad_norm": 4.8704705238342285,
+      "learning_rate": 0.00019983691039261357,
+      "loss": 1.6262,
       "step": 210
     },
     {
-      "epoch": 2.689655172413793,
-      "grad_norm": 12.651631355285645,
-      "learning_rate": 0.00019211662979500468,
-      "loss": 0.1925,
+      "epoch": 0.5721889554224884,
+      "grad_norm": 6.425934314727783,
+      "learning_rate": 0.0001998246095620893,
+      "loss": 1.4669,
       "step": 215
     },
     {
-      "epoch": 2.7523510971786833,
-      "grad_norm": 14.720685958862305,
-      "learning_rate": 0.00019172113015054532,
-      "loss": 0.3322,
+      "epoch": 0.5854956753160346,
+      "grad_norm": 9.25566291809082,
+      "learning_rate": 0.0001998118619612634,
+      "loss": 1.6061,
       "step": 220
     },
     {
-      "epoch": 2.815047021943574,
-      "grad_norm": 15.48449993133545,
-      "learning_rate": 0.00019131637858322225,
-      "loss": 0.1582,
+      "epoch": 0.5988023952095808,
+      "grad_norm": 3.8261756896972656,
+      "learning_rate": 0.00019979866764718843,
+      "loss": 1.4114,
       "step": 225
     },
     {
-      "epoch": 2.877742946708464,
-      "grad_norm": 8.015260696411133,
-      "learning_rate": 0.00019090241592038113,
-      "loss": 0.3334,
+      "epoch": 0.612109115103127,
+      "grad_norm": 7.696674346923828,
+      "learning_rate": 0.00019978502667891625,
+      "loss": 1.8654,
       "step": 230
     },
     {
-      "epoch": 2.9404388714733543,
-      "grad_norm": 20.96394157409668,
-      "learning_rate": 0.00019047928391849195,
-      "loss": 0.2362,
+      "epoch": 0.6254158349966733,
+      "grad_norm": 7.678075313568115,
+      "learning_rate": 0.0001997709391174977,
+      "loss": 1.6465,
       "step": 235
     },
     {
-      "epoch": 3.0,
-      "grad_norm": 37.81004333496094,
-      "learning_rate": 0.00019004702525893732,
-      "loss": 0.2647,
-      "step": 240
-    },
-    {
-      "epoch": 3.0,
-      "eval_loss": 1.3779538869857788,
-      "eval_macro_f1": 73.11543717930599,
-      "eval_macro_precision": 73.228985585862,
-      "eval_macro_recall": 73.23809523809524,
-      "eval_micro_f1": 73.23809523809524,
-      "eval_micro_precision": 73.23809523809524,
-      "eval_micro_recall": 73.23809523809524,
-      "eval_runtime": 3.6271,
-      "eval_samples_per_second": 289.487,
-      "eval_steps_per_second": 18.196,
+      "epoch": 0.6387225548902196,
+      "grad_norm": 6.352901458740234,
+      "learning_rate": 0.00019975640502598244,
+      "loss": 1.3295,
       "step": 240
     },
     {
-      "epoch": 3.06269592476489,
-      "grad_norm": 8.566853523254395,
-      "learning_rate": 0.0001896056835437069,
-      "loss": 0.0999,
+      "epoch": 0.6520292747837658,
+      "grad_norm": 5.327084541320801,
+      "learning_rate": 0.00019974142446941851,
+      "loss": 1.496,
       "step": 245
     },
     {
-      "epoch": 3.1253918495297803,
-      "grad_norm": 8.521112442016602,
-      "learning_rate": 0.0001891553032909996,
-      "loss": 0.0719,
+      "epoch": 0.6653359946773121,
+      "grad_norm": 6.817936897277832,
+      "learning_rate": 0.00019972599751485226,
+      "loss": 1.3759,
       "step": 250
     },
     {
-      "epoch": 3.188087774294671,
-      "grad_norm": 16.376962661743164,
-      "learning_rate": 0.00018869592993073258,
-      "loss": 0.2835,
+      "epoch": 0.6786427145708582,
+      "grad_norm": 3.684908628463745,
+      "learning_rate": 0.00019971012423132775,
+      "loss": 0.9475,
       "step": 255
     },
     {
-      "epoch": 3.250783699059561,
-      "grad_norm": 33.38531494140625,
-      "learning_rate": 0.0001882276097999592,
-      "loss": 0.2702,
+      "epoch": 0.6919494344644045,
+      "grad_norm": 5.260998725891113,
+      "learning_rate": 0.00019969380468988677,
+      "loss": 1.0685,
       "step": 260
     },
     {
-      "epoch": 3.3134796238244513,
-      "grad_norm": 24.302696228027344,
-      "learning_rate": 0.0001877503901381947,
-      "loss": 0.2763,
+      "epoch": 0.7052561543579507,
+      "grad_norm": 8.901612281799316,
+      "learning_rate": 0.00019967703896356823,
+      "loss": 1.4494,
       "step": 265
     },
     {
-      "epoch": 3.376175548589342,
-      "grad_norm": 18.85578727722168,
-      "learning_rate": 0.0001872643190826512,
-      "loss": 0.5174,
+      "epoch": 0.718562874251497,
+      "grad_norm": 6.55958890914917,
+      "learning_rate": 0.00019965982712740808,
+      "loss": 1.5111,
       "step": 270
     },
     {
-      "epoch": 3.438871473354232,
-      "grad_norm": 53.06370162963867,
-      "learning_rate": 0.00018676944566338213,
-      "loss": 0.2884,
+      "epoch": 0.7318695941450433,
+      "grad_norm": 5.0339202880859375,
+      "learning_rate": 0.00019964216925843874,
+      "loss": 1.2905,
       "step": 275
     },
     {
-      "epoch": 3.501567398119122,
-      "grad_norm": 34.46812438964844,
-      "learning_rate": 0.0001862658197983366,
-      "loss": 0.2232,
+      "epoch": 0.7451763140385895,
+      "grad_norm": 7.584630966186523,
+      "learning_rate": 0.00019962406543568898,
+      "loss": 1.2245,
       "step": 280
     },
     {
-      "epoch": 3.5642633228840124,
-      "grad_norm": 18.52553939819336,
-      "learning_rate": 0.00018575349228832397,
-      "loss": 0.1158,
+      "epoch": 0.7584830339321357,
+      "grad_norm": 10.351391792297363,
+      "learning_rate": 0.0001996055157401834,
+      "loss": 1.407,
       "step": 285
     },
     {
-      "epoch": 3.626959247648903,
-      "grad_norm": 0.9173059463500977,
-      "learning_rate": 0.00018523251481188986,
-      "loss": 0.2123,
+      "epoch": 0.7717897538256819,
+      "grad_norm": 6.787538051605225,
+      "learning_rate": 0.00019958652025494212,
+      "loss": 1.718,
       "step": 290
     },
     {
-      "epoch": 3.689655172413793,
-      "grad_norm": 12.346527099609375,
-      "learning_rate": 0.00018470293992010312,
-      "loss": 0.2126,
+      "epoch": 0.7850964737192282,
+      "grad_norm": 4.587769508361816,
+      "learning_rate": 0.00019956707906498044,
+      "loss": 1.5187,
       "step": 295
     },
     {
-      "epoch": 3.7523510971786833,
-      "grad_norm": 84.08663177490234,
-      "learning_rate": 0.00018416482103125506,
-      "loss": 0.222,
+      "epoch": 0.7984031936127745,
+      "grad_norm": 6.607381820678711,
+      "learning_rate": 0.00019954719225730847,
+      "loss": 1.4279,
       "step": 300
     },
     {
-      "epoch": 3.815047021943574,
-      "grad_norm": 19.547080993652344,
-      "learning_rate": 0.0001836182124254711,
-      "loss": 0.1389,
+      "epoch": 0.8117099135063207,
+      "grad_norm": 9.418729782104492,
+      "learning_rate": 0.00019952685992093063,
+      "loss": 1.3854,
       "step": 305
     },
     {
-      "epoch": 3.877742946708464,
-      "grad_norm": 1.2908215522766113,
-      "learning_rate": 0.00018306316923923563,
-      "loss": 0.1632,
+      "epoch": 0.825016633399867,
+      "grad_norm": 7.311931133270264,
+      "learning_rate": 0.00019950608214684535,
+      "loss": 1.5504,
       "step": 310
     },
     {
-      "epoch": 3.9404388714733543,
-      "grad_norm": 30.96961784362793,
-      "learning_rate": 0.00018249974745983023,
-      "loss": 0.2089,
+      "epoch": 0.8383233532934131,
+      "grad_norm": 7.015618324279785,
+      "learning_rate": 0.0001994848590280447,
+      "loss": 1.528,
       "step": 315
     },
     {
-      "epoch": 4.0,
-      "grad_norm": 13.291217803955078,
-      "learning_rate": 0.00018192800391968642,
-      "loss": 0.3228,
-      "step": 320
-    },
-    {
-      "epoch": 4.0,
-      "eval_loss": 1.7461718320846558,
-      "eval_macro_f1": 72.72572059706258,
-      "eval_macro_precision": 72.71473132903009,
-      "eval_macro_recall": 72.76190476190476,
-      "eval_micro_f1": 72.76190476190476,
-      "eval_micro_precision": 72.76190476190476,
-      "eval_micro_recall": 72.76190476190476,
-      "eval_runtime": 4.7844,
-      "eval_samples_per_second": 219.463,
-      "eval_steps_per_second": 13.795,
+      "epoch": 0.8516300731869594,
+      "grad_norm": 5.8235697746276855,
+      "learning_rate": 0.00019946319065951382,
+      "loss": 1.0753,
       "step": 320
     },
     {
-      "epoch": 4.06269592476489,
-      "grad_norm": 22.18653678894043,
-      "learning_rate": 0.00018134799629065276,
-      "loss": 0.3768,
+      "epoch": 0.8649367930805056,
+      "grad_norm": 4.4757513999938965,
+      "learning_rate": 0.0001994410771382307,
+      "loss": 1.4042,
       "step": 325
     },
     {
-      "epoch": 4.12539184952978,
-      "grad_norm": 18.046934127807617,
-      "learning_rate": 0.00018075978307817764,
-      "loss": 0.1879,
+      "epoch": 0.8782435129740519,
+      "grad_norm": 6.9428510665893555,
+      "learning_rate": 0.00019941851856316548,
+      "loss": 1.3328,
       "step": 330
     },
     {
-      "epoch": 4.1880877742946705,
-      "grad_norm": 8.421941757202148,
-      "learning_rate": 0.00018016342361540773,
-      "loss": 0.2067,
+      "epoch": 0.8915502328675982,
+      "grad_norm": 5.409964084625244,
+      "learning_rate": 0.0001993955150352803,
+      "loss": 1.2911,
       "step": 335
     },
     {
-      "epoch": 4.250783699059561,
-      "grad_norm": 37.9600830078125,
-      "learning_rate": 0.0001795589780572031,
-      "loss": 0.1612,
+      "epoch": 0.9048569527611444,
+      "grad_norm": 6.9928297996521,
+      "learning_rate": 0.0001993720666575287,
+      "loss": 1.379,
       "step": 340
     },
     {
-      "epoch": 4.313479623824452,
-      "grad_norm": 2.552628993988037,
-      "learning_rate": 0.0001789465073740693,
-      "loss": 0.0256,
+      "epoch": 0.9181636726546906,
+      "grad_norm": 5.91607666015625,
+      "learning_rate": 0.00019934817353485501,
+      "loss": 1.3604,
       "step": 345
     },
     {
-      "epoch": 4.376175548589342,
-      "grad_norm": 45.092559814453125,
-      "learning_rate": 0.00017832607334600746,
-      "loss": 0.2179,
+      "epoch": 0.9314703925482368,
+      "grad_norm": 6.265745162963867,
+      "learning_rate": 0.00019932383577419432,
+      "loss": 1.3839,
       "step": 350
     },
     {
-      "epoch": 4.438871473354232,
-      "grad_norm": 1.6102479696273804,
-      "learning_rate": 0.00017769773855628226,
-      "loss": 0.1914,
+      "epoch": 0.9447771124417831,
+      "grad_norm": 6.445305824279785,
+      "learning_rate": 0.0001992990534844714,
+      "loss": 1.3727,
       "step": 355
     },
     {
-      "epoch": 4.501567398119122,
-      "grad_norm": 27.624589920043945,
-      "learning_rate": 0.0001770615663851093,
-      "loss": 0.1637,
+      "epoch": 0.9580838323353293,
+      "grad_norm": 8.271346092224121,
+      "learning_rate": 0.00019927382677660088,
+      "loss": 1.4262,
       "step": 360
     },
     {
-      "epoch": 4.564263322884012,
-      "grad_norm": 31.627342224121094,
-      "learning_rate": 0.00017641762100326208,
-      "loss": 0.1678,
+      "epoch": 0.9713905522288756,
+      "grad_norm": 5.764677047729492,
+      "learning_rate": 0.00019924815576348603,
+      "loss": 1.4416,
       "step": 365
     },
     {
-      "epoch": 4.6269592476489025,
-      "grad_norm": 17.774938583374023,
-      "learning_rate": 0.0001757659673655986,
-      "loss": 0.0717,
+      "epoch": 0.9846972721224219,
+      "grad_norm": 7.0903215408325195,
+      "learning_rate": 0.00019922204056001895,
+      "loss": 1.3508,
       "step": 370
     },
     {
-      "epoch": 4.689655172413794,
-      "grad_norm": 5.605461597442627,
-      "learning_rate": 0.00017510667120450998,
-      "loss": 0.2968,
+      "epoch": 0.998003992015968,
+      "grad_norm": 8.942666053771973,
+      "learning_rate": 0.00019919548128307954,
+      "loss": 1.2008,
       "step": 375
     },
     {
-      "epoch": 4.752351097178684,
-      "grad_norm": 27.00714683532715,
-      "learning_rate": 0.00017443979902328956,
-      "loss": 0.5085,
+      "epoch": 0.998003992015968,
+      "eval_loss": 0.37667316198349,
+      "eval_macro_f1": 59.54495855355076,
+      "eval_macro_precision": 76.51384999324237,
+      "eval_macro_recall": 54.28674905628468,
+      "eval_micro_f1": 86.71407185628742,
+      "eval_micro_precision": 86.71407185628742,
+      "eval_micro_recall": 86.71407185628742,
+      "eval_runtime": 6.2328,
+      "eval_samples_per_second": 428.701,
+      "eval_steps_per_second": 26.794,
+      "step": 375
+    },
+    {
+      "epoch": 1.0133067198935461,
+      "grad_norm": 4.265635967254639,
+      "learning_rate": 0.00019916847805153526,
+      "loss": 1.0752,
       "step": 380
     },
     {
-      "epoch": 4.815047021943574,
-      "grad_norm": 26.765687942504883,
-      "learning_rate": 0.0001737654180894249,
-      "loss": 0.1005,
+      "epoch": 1.0266134397870925,
+      "grad_norm": 5.709604740142822,
+      "learning_rate": 0.00019914103098624055,
+      "loss": 0.7603,
       "step": 385
     },
     {
-      "epoch": 4.877742946708464,
-      "grad_norm": 15.505696296691895,
-      "learning_rate": 0.00017308359642781242,
-      "loss": 0.197,
+      "epoch": 1.0399201596806387,
+      "grad_norm": 6.586650848388672,
+      "learning_rate": 0.00019911314021003613,
+      "loss": 0.7114,
       "step": 390
     },
     {
-      "epoch": 4.940438871473354,
-      "grad_norm": 7.485856056213379,
-      "learning_rate": 0.00017239440281389582,
-      "loss": 0.1499,
+      "epoch": 1.053226879574185,
+      "grad_norm": 3.8452811241149902,
+      "learning_rate": 0.00019908480584774867,
+      "loss": 0.4094,
       "step": 395
     },
     {
-      "epoch": 5.0,
-      "grad_norm": 2.3405601978302,
-      "learning_rate": 0.00017169790676672858,
-      "loss": 0.0747,
+      "epoch": 1.0665335994677312,
+      "grad_norm": 23.499847412109375,
+      "learning_rate": 0.00019905602802619007,
+      "loss": 0.7433,
       "step": 400
     },
     {
-      "epoch": 5.0,
-      "eval_loss": 1.7594363689422607,
-      "eval_macro_f1": 74.17047440340497,
-      "eval_macro_precision": 75.04188054945206,
-      "eval_macro_recall": 73.99999999999999,
-      "eval_micro_f1": 74.0,
-      "eval_micro_precision": 74.0,
-      "eval_micro_recall": 74.0,
-      "eval_runtime": 4.2956,
-      "eval_samples_per_second": 244.435,
-      "eval_steps_per_second": 15.364,
-      "step": 400
-    },
-    {
-      "epoch": 5.06269592476489,
-      "grad_norm": 11.809943199157715,
-      "learning_rate": 0.00017099417854196165,
-      "loss": 0.0426,
+      "epoch": 1.0798403193612773,
+      "grad_norm": 10.368353843688965,
+      "learning_rate": 0.00019902680687415705,
+      "loss": 1.1225,
       "step": 405
     },
     {
-      "epoch": 5.12539184952978,
-      "grad_norm": 4.720785140991211,
-      "learning_rate": 0.00017028328912475668,
-      "loss": 0.044,
+      "epoch": 1.0931470392548237,
+      "grad_norm": 10.001039505004883,
+      "learning_rate": 0.00019899714252243035,
+      "loss": 0.7539,
       "step": 410
     },
     {
-      "epoch": 5.1880877742946705,
-      "grad_norm": 14.604973793029785,
-      "learning_rate": 0.00016956531022262585,
-      "loss": 0.1949,
+      "epoch": 1.1064537591483699,
+      "grad_norm": 6.002054214477539,
+      "learning_rate": 0.00019896703510377436,
+      "loss": 0.9843,
       "step": 415
     },
     {
-      "epoch": 5.250783699059561,
-      "grad_norm": 0.6709997653961182,
-      "learning_rate": 0.00016884031425819853,
-      "loss": 0.295,
+      "epoch": 1.1197604790419162,
+      "grad_norm": 14.071710586547852,
+      "learning_rate": 0.00019893648475293648,
+      "loss": 1.1239,
       "step": 420
     },
     {
-      "epoch": 5.313479623824452,
-      "grad_norm": 69.35284423828125,
-      "learning_rate": 0.00016810837436191623,
-      "loss": 0.4807,
+      "epoch": 1.1330671989354624,
+      "grad_norm": 4.876122951507568,
+      "learning_rate": 0.00019890549160664632,
+      "loss": 0.8278,
       "step": 425
     },
     {
-      "epoch": 5.376175548589342,
-      "grad_norm": 24.146665573120117,
-      "learning_rate": 0.00016736956436465573,
-      "loss": 0.5015,
+      "epoch": 1.1463739188290087,
+      "grad_norm": 7.125596046447754,
+      "learning_rate": 0.00019887405580361537,
+      "loss": 0.8853,
       "step": 430
     },
     {
-      "epoch": 5.438871473354232,
-      "grad_norm": 3.2245795726776123,
-      "learning_rate": 0.0001666239587902819,
-      "loss": 0.1359,
+      "epoch": 1.159680638722555,
+      "grad_norm": 4.5107645988464355,
+      "learning_rate": 0.00019884217748453623,
+      "loss": 0.7046,
       "step": 435
     },
     {
-      "epoch": 5.501567398119122,
-      "grad_norm": 8.290742874145508,
-      "learning_rate": 0.00016587163284813032,
-      "loss": 0.0412,
+      "epoch": 1.172987358616101,
+      "grad_norm": 7.978573322296143,
+      "learning_rate": 0.00019880985679208207,
+      "loss": 0.9162,
       "step": 440
     },
     {
-      "epoch": 5.564263322884012,
-      "grad_norm": 21.256576538085938,
-      "learning_rate": 0.00016511266242542102,
-      "loss": 0.0626,
+      "epoch": 1.1862940785096474,
+      "grad_norm": 11.903189659118652,
+      "learning_rate": 0.0001987770938709057,
+      "loss": 0.9984,
       "step": 445
     },
     {
-      "epoch": 5.6269592476489025,
-      "grad_norm": 1.178815245628357,
-      "learning_rate": 0.00016434712407960373,
-      "loss": 0.0422,
+      "epoch": 1.1996007984031936,
+      "grad_norm": 15.054693222045898,
+      "learning_rate": 0.00019874388886763944,
+      "loss": 0.9147,
       "step": 450
     },
     {
-      "epoch": 5.689655172413794,
-      "grad_norm": 19.474632263183594,
-      "learning_rate": 0.00016357509503063538,
-      "loss": 0.3287,
+      "epoch": 1.21290751829674,
+      "grad_norm": 9.752163887023926,
+      "learning_rate": 0.000198710241930894,
+      "loss": 0.868,
       "step": 455
     },
     {
-      "epoch": 5.752351097178684,
-      "grad_norm": 38.11343002319336,
-      "learning_rate": 0.00016279665315319114,
-      "loss": 0.3171,
+      "epoch": 1.226214238190286,
+      "grad_norm": 8.092059135437012,
+      "learning_rate": 0.00019867615321125795,
+      "loss": 0.7435,
       "step": 460
     },
     {
-      "epoch": 5.815047021943574,
-      "grad_norm": 63.41005325317383,
-      "learning_rate": 0.00016201187696880887,
-      "loss": 0.2117,
+      "epoch": 1.2395209580838324,
+      "grad_norm": 8.409940719604492,
+      "learning_rate": 0.0001986416228612972,
+      "loss": 0.7041,
       "step": 465
     },
     {
-      "epoch": 5.877742946708464,
-      "grad_norm": 1.7413610219955444,
-      "learning_rate": 0.00016122084563796905,
-      "loss": 0.0836,
+      "epoch": 1.2528276779773786,
+      "grad_norm": 6.191831588745117,
+      "learning_rate": 0.00019860665103555415,
+      "loss": 0.7856,
       "step": 470
     },
     {
-      "epoch": 5.940438871473354,
-      "grad_norm": 25.483036041259766,
-      "learning_rate": 0.00016042363895210946,
-      "loss": 0.4536,
+      "epoch": 1.2661343978709247,
+      "grad_norm": 11.023136138916016,
+      "learning_rate": 0.00019857123789054706,
+      "loss": 0.9971,
       "step": 475
     },
     {
-      "epoch": 6.0,
-      "grad_norm": 13.102269172668457,
-      "learning_rate": 0.00015962033732557686,
-      "loss": 0.0868,
-      "step": 480
-    },
-    {
-      "epoch": 6.0,
-      "eval_loss": 1.5808621644973755,
-      "eval_macro_f1": 73.24384906714903,
-      "eval_macro_precision": 73.61945929207427,
-      "eval_macro_recall": 73.61904761904762,
-      "eval_micro_f1": 73.61904761904762,
-      "eval_micro_precision": 73.61904761904762,
-      "eval_micro_recall": 73.61904761904762,
-      "eval_runtime": 5.0263,
-      "eval_samples_per_second": 208.902,
-      "eval_steps_per_second": 13.131,
+      "epoch": 1.2794411177644711,
+      "grad_norm": 12.192377090454102,
+      "learning_rate": 0.00019853538358476932,
+      "loss": 0.9177,
       "step": 480
     },
     {
-      "epoch": 6.06269592476489,
-      "grad_norm": 6.51555871963501,
-      "learning_rate": 0.00015881102178751553,
-      "loss": 0.0678,
+      "epoch": 1.2927478376580173,
+      "grad_norm": 6.031184673309326,
+      "learning_rate": 0.00019849908827868876,
+      "loss": 0.5903,
       "step": 485
     },
     {
-      "epoch": 6.12539184952978,
-      "grad_norm": 19.54594612121582,
-      "learning_rate": 0.00015799577397369375,
-      "loss": 0.0731,
+      "epoch": 1.3060545575515636,
+      "grad_norm": 5.597287654876709,
+      "learning_rate": 0.00019846235213474692,
+      "loss": 0.8199,
       "step": 490
     },
     {
-      "epoch": 6.1880877742946705,
-      "grad_norm": 7.0642266273498535,
-      "learning_rate": 0.0001571746761182693,
-      "loss": 0.0848,
+      "epoch": 1.3193612774451098,
+      "grad_norm": 12.112035751342773,
+      "learning_rate": 0.00019842517531735838,
+      "loss": 0.976,
       "step": 495
     },
     {
-      "epoch": 6.250783699059561,
-      "grad_norm": 4.012495040893555,
-      "learning_rate": 0.00015634781104549442,
-      "loss": 0.1141,
+      "epoch": 1.332667997338656,
+      "grad_norm": 8.983979225158691,
+      "learning_rate": 0.00019838755799290994,
+      "loss": 0.6765,
       "step": 500
     },
     {
-      "epoch": 6.313479623824452,
-      "grad_norm": 2.1113383769989014,
-      "learning_rate": 0.00015551526216136144,
-      "loss": 0.0945,
+      "epoch": 1.3459747172322023,
+      "grad_norm": 11.685017585754395,
+      "learning_rate": 0.00019834950032975986,
+      "loss": 0.7698,
       "step": 505
     },
     {
-      "epoch": 6.376175548589342,
-      "grad_norm": 37.096473693847656,
-      "learning_rate": 0.00015467711344518942,
-      "loss": 0.225,
+      "epoch": 1.3592814371257484,
+      "grad_norm": 7.77982759475708,
+      "learning_rate": 0.00019831100249823733,
+      "loss": 0.7094,
       "step": 510
     },
     {
-      "epoch": 6.438871473354232,
-      "grad_norm": 25.69189453125,
-      "learning_rate": 0.00015383344944115324,
-      "loss": 0.0556,
+      "epoch": 1.3725881570192948,
+      "grad_norm": 7.004415035247803,
+      "learning_rate": 0.00019827206467064133,
+      "loss": 0.8004,
       "step": 515
     },
     {
-      "epoch": 6.501567398119122,
-      "grad_norm": 4.062742710113525,
-      "learning_rate": 0.00015298435524975572,
-      "loss": 0.0765,
+      "epoch": 1.385894876912841,
+      "grad_norm": 9.52391242980957,
+      "learning_rate": 0.0001982326870212402,
+      "loss": 0.7738,
       "step": 520
     },
     {
-      "epoch": 6.564263322884012,
-      "grad_norm": 14.947588920593262,
-      "learning_rate": 0.00015212991651924326,
-      "loss": 0.0302,
+      "epoch": 1.3992015968063871,
+      "grad_norm": 6.2501726150512695,
+      "learning_rate": 0.00019819286972627066,
+      "loss": 0.7728,
       "step": 525
     },
     {
-      "epoch": 6.6269592476489025,
-      "grad_norm": 28.04601287841797,
-      "learning_rate": 0.0001512702194369668,
-      "loss": 0.4446,
+      "epoch": 1.4125083166999335,
+      "grad_norm": 11.152785301208496,
+      "learning_rate": 0.00019815261296393715,
+      "loss": 0.946,
       "step": 530
     },
     {
-      "epoch": 6.689655172413794,
-      "grad_norm": 0.8309017419815063,
-      "learning_rate": 0.00015040535072068778,
-      "loss": 0.1228,
+      "epoch": 1.4258150365934796,
+      "grad_norm": 10.387774467468262,
+      "learning_rate": 0.0001981119169144109,
+      "loss": 0.6949,
       "step": 535
     },
     {
-      "epoch": 6.752351097178684,
-      "grad_norm": 0.3538524806499481,
-      "learning_rate": 0.00014953539760983122,
-      "loss": 0.1258,
+      "epoch": 1.439121756487026,
+      "grad_norm": 10.800827980041504,
+      "learning_rate": 0.00019807078175982924,
+      "loss": 0.9506,
       "step": 540
     },
     {
-      "epoch": 6.815047021943574,
-      "grad_norm": 0.2524064779281616,
-      "learning_rate": 0.00014866044785668563,
-      "loss": 0.036,
+      "epoch": 1.4524284763805722,
+      "grad_norm": 11.124509811401367,
+      "learning_rate": 0.0001980292076842947,
+      "loss": 1.139,
       "step": 545
     },
     {
-      "epoch": 6.877742946708464,
-      "grad_norm": 46.764183044433594,
-      "learning_rate": 0.00014778058971755154,
-      "loss": 0.294,
+      "epoch": 1.4657351962741183,
+      "grad_norm": 8.33337116241455,
+      "learning_rate": 0.00019798719487387428,
+      "loss": 0.8012,
       "step": 550
     },
     {
-      "epoch": 6.940438871473354,
-      "grad_norm": 33.56895446777344,
-      "learning_rate": 0.00014689591194383896,
-      "loss": 0.1452,
+      "epoch": 1.4790419161676647,
+      "grad_norm": 16.182756423950195,
+      "learning_rate": 0.00019794474351659852,
+      "loss": 0.7395,
       "step": 555
     },
     {
-      "epoch": 7.0,
-      "grad_norm": 0.07464821636676788,
-      "learning_rate": 0.00014600650377311522,
-      "loss": 0.0801,
-      "step": 560
-    },
-    {
-      "epoch": 7.0,
-      "eval_loss": 1.7074089050292969,
-      "eval_macro_f1": 74.3995628020244,
-      "eval_macro_precision": 74.43261914638138,
-      "eval_macro_recall": 74.38095238095238,
-      "eval_micro_f1": 74.38095238095238,
-      "eval_micro_precision": 74.38095238095238,
-      "eval_micro_recall": 74.38095238095238,
-      "eval_runtime": 5.0045,
-      "eval_samples_per_second": 209.812,
-      "eval_steps_per_second": 13.188,
+      "epoch": 1.492348636061211,
+      "grad_norm": 7.745892524719238,
+      "learning_rate": 0.0001979018538024607,
+      "loss": 0.7877,
       "step": 560
     },
     {
-      "epoch": 7.06269592476489,
-      "grad_norm": 24.801523208618164,
-      "learning_rate": 0.00014511245492010335,
-      "loss": 0.0224,
+      "epoch": 1.5056553559547572,
+      "grad_norm": 5.482974529266357,
+      "learning_rate": 0.00019785852592341598,
+      "loss": 0.7299,
       "step": 565
     },
     {
-      "epoch": 7.12539184952978,
-      "grad_norm": 2.976022720336914,
-      "learning_rate": 0.00014421385556763266,
-      "loss": 0.1211,
+      "epoch": 1.5189620758483033,
+      "grad_norm": 9.263693809509277,
+      "learning_rate": 0.00019781476007338058,
+      "loss": 0.9189,
       "step": 570
     },
     {
-      "epoch": 7.1880877742946705,
-      "grad_norm": 0.37286555767059326,
-      "learning_rate": 0.00014331079635754193,
-      "loss": 0.0327,
+      "epoch": 1.5322687957418495,
+      "grad_norm": 10.035237312316895,
+      "learning_rate": 0.00019777055644823086,
+      "loss": 0.7287,
       "step": 575
     },
     {
-      "epoch": 7.250783699059561,
-      "grad_norm": 0.7439866662025452,
-      "learning_rate": 0.0001424033683815365,
-      "loss": 0.0382,
+      "epoch": 1.5455755156353959,
+      "grad_norm": 6.058926105499268,
+      "learning_rate": 0.0001977259152458025,
+      "loss": 1.0947,
       "step": 580
     },
     {
-      "epoch": 7.313479623824452,
-      "grad_norm": 24.909921646118164,
-      "learning_rate": 0.00014149166317199954,
-      "loss": 0.0603,
+      "epoch": 1.5588822355289422,
+      "grad_norm": 11.625580787658691,
+      "learning_rate": 0.00019768083666588953,
+      "loss": 1.0071,
       "step": 585
     },
     {
-      "epoch": 7.376175548589342,
-      "grad_norm": 9.340410232543945,
-      "learning_rate": 0.0001405757726927595,
-      "loss": 0.2319,
+      "epoch": 1.5721889554224884,
+      "grad_norm": 6.705043315887451,
+      "learning_rate": 0.00019763532091024352,
+      "loss": 0.8665,
       "step": 590
     },
     {
-      "epoch": 7.438871473354232,
-      "grad_norm": 3.947765350341797,
-      "learning_rate": 0.00013965578932981346,
-      "loss": 0.0448,
+      "epoch": 1.5854956753160345,
+      "grad_norm": 5.264439582824707,
+      "learning_rate": 0.00019758936818257264,
+      "loss": 0.767,
       "step": 595
     },
     {
-      "epoch": 7.501567398119122,
-      "grad_norm": 0.29921436309814453,
-      "learning_rate": 0.00013873180588200827,
-      "loss": 0.006,
+      "epoch": 1.5988023952095807,
+      "grad_norm": 14.696879386901855,
+      "learning_rate": 0.00019754297868854073,
+      "loss": 0.628,
       "step": 600
     },
     {
-      "epoch": 7.564263322884012,
-      "grad_norm": 0.11227085441350937,
-      "learning_rate": 0.00013780391555167992,
-      "loss": 0.0037,
+      "epoch": 1.612109115103127,
+      "grad_norm": 10.865748405456543,
+      "learning_rate": 0.0001974961526357664,
+      "loss": 0.7197,
       "step": 605
     },
     {
-      "epoch": 7.6269592476489025,
-      "grad_norm": 0.041259847581386566,
-      "learning_rate": 0.0001368722119352521,
-      "loss": 0.0149,
+      "epoch": 1.6254158349966734,
+      "grad_norm": 8.621575355529785,
+      "learning_rate": 0.00019744889023382215,
+      "loss": 1.1014,
       "step": 610
     },
     {
-      "epoch": 7.689655172413794,
-      "grad_norm": 41.212154388427734,
-      "learning_rate": 0.00013593678901379524,
-      "loss": 0.0205,
+      "epoch": 1.6387225548902196,
+      "grad_norm": 11.162896156311035,
+      "learning_rate": 0.00019740119169423337,
+      "loss": 0.9153,
       "step": 615
     },
     {
-      "epoch": 7.752351097178684,
-      "grad_norm": 12.241500854492188,
-      "learning_rate": 0.00013499774114354655,
-      "loss": 0.0038,
+      "epoch": 1.6520292747837657,
+      "grad_norm": 10.98604679107666,
+      "learning_rate": 0.00019735305723047732,
+      "loss": 1.0044,
       "step": 620
     },
     {
-      "epoch": 7.815047021943574,
-      "grad_norm": 55.03153991699219,
-      "learning_rate": 0.00013405516304639234,
-      "loss": 0.029,
+      "epoch": 1.665335994677312,
+      "grad_norm": 10.798745155334473,
+      "learning_rate": 0.00019730448705798239,
+      "loss": 1.2465,
       "step": 625
     },
     {
-      "epoch": 7.877742946708464,
-      "grad_norm": 0.5877093076705933,
-      "learning_rate": 0.00013310914980031334,
-      "loss": 0.0072,
+      "epoch": 1.6786427145708582,
+      "grad_norm": 7.607357978820801,
+      "learning_rate": 0.00019725548139412692,
+      "loss": 0.9162,
       "step": 630
     },
     {
-      "epoch": 7.940438871473354,
-      "grad_norm": 34.05458068847656,
-      "learning_rate": 0.00013215979682979428,
-      "loss": 0.0398,
+      "epoch": 1.6919494344644046,
+      "grad_norm": 8.403728485107422,
+      "learning_rate": 0.00019720604045823836,
+      "loss": 0.8765,
       "step": 635
     },
     {
-      "epoch": 8.0,
-      "grad_norm": 86.24744415283203,
-      "learning_rate": 0.00013120719989619833,
-      "loss": 0.1489,
-      "step": 640
-    },
-    {
-      "epoch": 8.0,
-      "eval_loss": 3.0941507816314697,
-      "eval_macro_f1": 73.75591198712105,
-      "eval_macro_precision": 74.97115819170575,
-      "eval_macro_recall": 73.52380952380952,
-      "eval_micro_f1": 73.52380952380952,
-      "eval_micro_precision": 73.52380952380952,
-      "eval_micro_recall": 73.52380952380952,
-      "eval_runtime": 4.8846,
-      "eval_samples_per_second": 214.962,
-      "eval_steps_per_second": 13.512,
+      "epoch": 1.7052561543579507,
+      "grad_norm": 8.099462509155273,
+      "learning_rate": 0.0001971561644715922,
+      "loss": 0.9799,
       "step": 640
     },
     {
-      "epoch": 8.06269592476489,
-      "grad_norm": 0.29403501749038696,
-      "learning_rate": 0.0001302514550881076,
-      "loss": 0.1469,
+      "epoch": 1.718562874251497,
+      "grad_norm": 10.743642807006836,
+      "learning_rate": 0.00019710585365741103,
+      "loss": 0.8505,
       "step": 645
     },
     {
-      "epoch": 8.12539184952978,
-      "grad_norm": 0.0023457759525626898,
-      "learning_rate": 0.0001292926588116308,
-      "loss": 0.0521,
+      "epoch": 1.7318695941450433,
+      "grad_norm": 6.268372535705566,
+      "learning_rate": 0.0001970551082408636,
+      "loss": 0.6592,
       "step": 650
     },
     {
-      "epoch": 8.18808777429467,
-      "grad_norm": 0.0002847153227776289,
-      "learning_rate": 0.00012833090778067863,
-      "loss": 0.0004,
+      "epoch": 1.7451763140385896,
+      "grad_norm": 11.54825210571289,
+      "learning_rate": 0.0001970039284490637,
+      "loss": 1.1281,
       "step": 655
     },
     {
-      "epoch": 8.25078369905956,
-      "grad_norm": 0.07453305274248123,
-      "learning_rate": 0.0001273662990072083,
-      "loss": 0.0002,
+      "epoch": 1.7584830339321358,
+      "grad_norm": 8.397130012512207,
+      "learning_rate": 0.00019695231451106912,
+      "loss": 0.8406,
       "step": 660
     },
     {
-      "epoch": 8.31347962382445,
-      "grad_norm": 0.015363110229372978,
-      "learning_rate": 0.00012639892979143789,
-      "loss": 0.0349,
+      "epoch": 1.771789753825682,
+      "grad_norm": 6.930296897888184,
+      "learning_rate": 0.00019690026665788085,
+      "loss": 0.6492,
       "step": 665
     },
     {
-      "epoch": 8.376175548589341,
-      "grad_norm": 0.003746453206986189,
-      "learning_rate": 0.00012542889771203166,
-      "loss": 0.0002,
+      "epoch": 1.785096473719228,
+      "grad_norm": 8.116912841796875,
+      "learning_rate": 0.00019684778512244172,
+      "loss": 0.8172,
       "step": 670
     },
     {
-      "epoch": 8.438871473354231,
-      "grad_norm": 0.07255641371011734,
-      "learning_rate": 0.00012445630061625729,
-      "loss": 0.0391,
+      "epoch": 1.7984031936127745,
+      "grad_norm": 8.120738983154297,
+      "learning_rate": 0.00019679487013963564,
+      "loss": 0.7236,
       "step": 675
     },
     {
-      "epoch": 8.501567398119121,
-      "grad_norm": 0.006769082974642515,
-      "learning_rate": 0.00012348123661011601,
-      "loss": 0.0626,
+      "epoch": 1.8117099135063208,
+      "grad_norm": 9.556026458740234,
+      "learning_rate": 0.00019674152194628638,
+      "loss": 1.0562,
       "step": 680
     },
     {
-      "epoch": 8.564263322884013,
-      "grad_norm": 1.8504444360733032,
-      "learning_rate": 0.00012250380404844665,
-      "loss": 0.1377,
+      "epoch": 1.825016633399867,
+      "grad_norm": 8.71633529663086,
+      "learning_rate": 0.00019668774078115665,
+      "loss": 0.6465,
       "step": 685
     },
     {
-      "epoch": 8.626959247648903,
-      "grad_norm": 30.328672409057617,
-      "learning_rate": 0.00012152410152500453,
-      "loss": 0.0094,
+      "epoch": 1.8383233532934131,
+      "grad_norm": 7.23399019241333,
+      "learning_rate": 0.00019663352688494684,
+      "loss": 0.7309,
       "step": 690
     },
     {
-      "epoch": 8.689655172413794,
-      "grad_norm": 0.01078982837498188,
-      "learning_rate": 0.00012054222786251634,
-      "loss": 0.0014,
+      "epoch": 1.8516300731869593,
+      "grad_norm": 10.535300254821777,
+      "learning_rate": 0.00019657888050029414,
+      "loss": 0.891,
       "step": 695
     },
     {
-      "epoch": 8.752351097178684,
-      "grad_norm": 0.02426566183567047,
-      "learning_rate": 0.00011955828210271187,
-      "loss": 0.0001,
+      "epoch": 1.8649367930805056,
+      "grad_norm": 6.320178985595703,
+      "learning_rate": 0.00019652380187177126,
+      "loss": 0.8396,
       "step": 700
     },
     {
-      "epoch": 8.815047021943574,
-      "grad_norm": 0.3005918562412262,
-      "learning_rate": 0.00011857236349633358,
-      "loss": 0.1305,
+      "epoch": 1.878243512974052,
+      "grad_norm": 7.133821487426758,
+      "learning_rate": 0.0001964682912458856,
+      "loss": 0.8449,
       "step": 705
     },
     {
-      "epoch": 8.877742946708464,
-      "grad_norm": 40.52849578857422,
-      "learning_rate": 0.00011758457149312538,
-      "loss": 0.1226,
+      "epoch": 1.8915502328675982,
+      "grad_norm": 9.44013786315918,
+      "learning_rate": 0.00019641234887107778,
+      "loss": 0.7616,
       "step": 710
     },
     {
-      "epoch": 8.940438871473354,
-      "grad_norm": 0.03640785068273544,
-      "learning_rate": 0.0001165950057318008,
-      "loss": 0.0081,
+      "epoch": 1.9048569527611443,
+      "grad_norm": 9.079183578491211,
+      "learning_rate": 0.00019635597499772093,
+      "loss": 1.027,
       "step": 715
     },
     {
-      "epoch": 9.0,
-      "grad_norm": 0.8364064693450928,
-      "learning_rate": 0.00011560376602999272,
-      "loss": 0.1558,
-      "step": 720
-    },
-    {
-      "epoch": 9.0,
-      "eval_loss": 3.2611937522888184,
-      "eval_macro_f1": 72.56355683169605,
-      "eval_macro_precision": 74.47633152198945,
-      "eval_macro_recall": 72.3809523809524,
-      "eval_micro_f1": 72.38095238095238,
-      "eval_micro_precision": 72.38095238095238,
-      "eval_micro_recall": 72.38095238095238,
-      "eval_runtime": 4.5879,
-      "eval_samples_per_second": 228.86,
-      "eval_steps_per_second": 14.386,
+      "epoch": 1.9181636726546905,
+      "grad_norm": 6.522233486175537,
+      "learning_rate": 0.00019629916987811926,
+      "loss": 1.0346,
       "step": 720
     },
     {
-      "epoch": 9.06269592476489,
-      "grad_norm": 0.017020411789417267,
-      "learning_rate": 0.00011461095237418454,
-      "loss": 0.0155,
+      "epoch": 1.9314703925482368,
+      "grad_norm": 12.074502944946289,
+      "learning_rate": 0.0001962419337665071,
+      "loss": 1.1165,
       "step": 725
     },
     {
-      "epoch": 9.12539184952978,
-      "grad_norm": 0.0012181233614683151,
-      "learning_rate": 0.00011361666490962468,
-      "loss": 0.0176,
+      "epoch": 1.9447771124417832,
+      "grad_norm": 6.275933265686035,
+      "learning_rate": 0.00019618426691904762,
+      "loss": 0.7029,
       "step": 730
     },
     {
-      "epoch": 9.18808777429467,
-      "grad_norm": 0.04490913078188896,
-      "learning_rate": 0.00011262100393022482,
-      "loss": 0.0084,
+      "epoch": 1.9580838323353293,
+      "grad_norm": 9.753368377685547,
+      "learning_rate": 0.0001961261695938319,
+      "loss": 0.9627,
       "step": 735
     },
     {
-      "epoch": 9.25078369905956,
-      "grad_norm": 0.0009038946591317654,
-      "learning_rate": 0.00011162406986844323,
-      "loss": 0.0,
+      "epoch": 1.9713905522288755,
+      "grad_norm": 8.683499336242676,
+      "learning_rate": 0.00019606764205087757,
+      "loss": 0.7202,
       "step": 740
     },
     {
-      "epoch": 9.31347962382445,
-      "grad_norm": 0.0007440209737978876,
-      "learning_rate": 0.00011062596328515424,
-      "loss": 0.0207,
+      "epoch": 1.9846972721224219,
+      "grad_norm": 8.230753898620605,
+      "learning_rate": 0.0001960086845521277,
+      "loss": 0.8854,
       "step": 745
     },
     {
-      "epoch": 9.376175548589341,
-      "grad_norm": 0.00041356749716214836,
-      "learning_rate": 0.00010962678485950455,
-      "loss": 0.0018,
+      "epoch": 1.998003992015968,
+      "grad_norm": 8.35472583770752,
+      "learning_rate": 0.00019594929736144976,
+      "loss": 1.194,
       "step": 750
     },
     {
-      "epoch": 9.438871473354231,
-      "grad_norm": 0.0054762535728514194,
-      "learning_rate": 0.00010862663537875775,
-      "loss": 0.0,
+      "epoch": 1.998003992015968,
+      "eval_loss": 0.36028650403022766,
+      "eval_macro_f1": 68.21512864198317,
+      "eval_macro_precision": 70.69761046501635,
+      "eval_macro_recall": 66.15824103590107,
+      "eval_micro_f1": 87.3502994011976,
+      "eval_micro_precision": 87.3502994011976,
+      "eval_micro_recall": 87.3502994011976,
+      "eval_runtime": 5.244,
+      "eval_samples_per_second": 509.537,
+      "eval_steps_per_second": 31.846,
+      "step": 750
+    },
+    {
+      "epoch": 2.013306719893546,
+      "grad_norm": 3.9793150424957275,
+      "learning_rate": 0.00019588948074463416,
+      "loss": 0.4671,
       "step": 755
     },
     {
-      "epoch": 9.501567398119121,
-      "grad_norm": 0.027165431529283524,
-      "learning_rate": 0.00010762561572812788,
-      "loss": 0.0,
+      "epoch": 2.0266134397870923,
+      "grad_norm": 5.490118503570557,
+      "learning_rate": 0.00019582923496939337,
+      "loss": 0.4299,
       "step": 760
     },
     {
-      "epoch": 9.564263322884013,
-      "grad_norm": 0.002276531420648098,
-      "learning_rate": 0.0001066238268806032,
-      "loss": 0.0102,
+      "epoch": 2.039920159680639,
+      "grad_norm": 11.58049201965332,
+      "learning_rate": 0.00019576856030536054,
+      "loss": 0.5549,
       "step": 765
     },
     {
-      "epoch": 9.626959247648903,
-      "grad_norm": 0.14534221589565277,
-      "learning_rate": 0.00010562136988676078,
-      "loss": 0.0001,
+      "epoch": 2.053226879574185,
+      "grad_norm": 12.648699760437012,
+      "learning_rate": 0.0001957074570240883,
+      "loss": 0.2281,
       "step": 770
     },
     {
-      "epoch": 9.689655172413794,
-      "grad_norm": 0.0010948021663352847,
-      "learning_rate": 0.00010461834586457398,
-      "loss": 0.0,
+      "epoch": 2.066533599467731,
+      "grad_norm": 7.2631754875183105,
+      "learning_rate": 0.0001956459253990476,
+      "loss": 0.35,
       "step": 775
     },
     {
-      "epoch": 9.752351097178684,
-      "grad_norm": 5.5901953601278365e-05,
-      "learning_rate": 0.00010361485598921212,
-      "loss": 0.104,
+      "epoch": 2.0798403193612773,
+      "grad_norm": 8.948901176452637,
+      "learning_rate": 0.0001955839657056265,
+      "loss": 0.3369,
       "step": 780
     },
     {
-      "epoch": 9.815047021943574,
-      "grad_norm": 0.8184991478919983,
-      "learning_rate": 0.00010261100148283538,
-      "loss": 0.0106,
+      "epoch": 2.0931470392548235,
+      "grad_norm": 4.452752113342285,
+      "learning_rate": 0.0001955215782211289,
+      "loss": 0.3592,
       "step": 785
     },
     {
-      "epoch": 9.877742946708464,
-      "grad_norm": 0.006693361327052116,
-      "learning_rate": 0.00010160688360438419,
-      "loss": 0.0,
+      "epoch": 2.10645375914837,
+      "grad_norm": 13.342286109924316,
+      "learning_rate": 0.0001954587632247732,
+      "loss": 0.5003,
       "step": 790
     },
     {
-      "epoch": 9.940438871473354,
-      "grad_norm": 59.32970428466797,
-      "learning_rate": 0.00010060260363936547,
-      "loss": 0.052,
+      "epoch": 2.1197604790419162,
+      "grad_norm": 5.800313472747803,
+      "learning_rate": 0.00019539552099769126,
+      "loss": 0.384,
       "step": 795
     },
     {
-      "epoch": 10.0,
-      "grad_norm": 0.021533485502004623,
-      "learning_rate": 9.959826288963571e-05,
-      "loss": 0.0,
-      "step": 800
-    },
-    {
-      "epoch": 10.0,
-      "eval_loss": 3.001224994659424,
-      "eval_macro_f1": 74.98933185346014,
-      "eval_macro_precision": 75.36484226701617,
-      "eval_macro_recall": 74.85714285714286,
-      "eval_micro_f1": 74.85714285714286,
-      "eval_micro_precision": 74.85714285714286,
-      "eval_micro_recall": 74.85714285714286,
-      "eval_runtime": 4.9316,
-      "eval_samples_per_second": 212.912,
-      "eval_steps_per_second": 13.383,
+      "epoch": 2.1330671989354624,
+      "grad_norm": 5.854800701141357,
+      "learning_rate": 0.00019533185182292703,
+      "loss": 0.3074,
       "step": 800
     },
     {
-      "epoch": 10.06269592476489,
-      "grad_norm": 0.10546494275331497,
-      "learning_rate": 9.85939626631829e-05,
-      "loss": 0.0001,
+      "epoch": 2.1463739188290085,
+      "grad_norm": 6.107956409454346,
+      "learning_rate": 0.00019526775598543522,
+      "loss": 0.3026,
       "step": 805
     },
     {
-      "epoch": 10.12539184952978,
-      "grad_norm": 0.0011004558764398098,
-      "learning_rate": 9.758980426390732e-05,
-      "loss": 0.0,
+      "epoch": 2.1596806387225547,
+      "grad_norm": 5.478348731994629,
+      "learning_rate": 0.00019520323377208017,
+      "loss": 0.3378,
       "step": 810
     },
     {
-      "epoch": 10.18808777429467,
-      "grad_norm": 0.02645757608115673,
-      "learning_rate": 9.658588898140322e-05,
-      "loss": 0.0,
+      "epoch": 2.1729873586161013,
+      "grad_norm": 8.135194778442383,
+      "learning_rate": 0.00019513828547163437,
+      "loss": 0.458,
       "step": 815
     },
     {
-      "epoch": 10.25078369905956,
-      "grad_norm": 0.06658319383859634,
-      "learning_rate": 9.558231808074156e-05,
-      "loss": 0.0001,
+      "epoch": 2.1862940785096474,
+      "grad_norm": 3.8288862705230713,
+      "learning_rate": 0.00019507291137477742,
+      "loss": 0.2756,
       "step": 820
     },
     {
-      "epoch": 10.31347962382445,
-      "grad_norm": 0.1821487843990326,
-      "learning_rate": 9.457919279225548e-05,
-      "loss": 0.0,
+      "epoch": 2.1996007984031936,
+      "grad_norm": 9.290178298950195,
+      "learning_rate": 0.00019500711177409454,
+      "loss": 0.2866,
       "step": 825
     },
     {
-      "epoch": 10.376175548589341,
-      "grad_norm": 0.0002453463093843311,
-      "learning_rate": 9.357661430132915e-05,
-      "loss": 0.0,
+      "epoch": 2.2129075182967397,
+      "grad_norm": 5.381500720977783,
+      "learning_rate": 0.00019494088696407532,
+      "loss": 0.2167,
       "step": 830
     },
     {
-      "epoch": 10.438871473354231,
-      "grad_norm": 0.0003151444543618709,
-      "learning_rate": 9.257468373819123e-05,
-      "loss": 0.0,
+      "epoch": 2.2262142381902863,
+      "grad_norm": 6.5513458251953125,
+      "learning_rate": 0.0001948742372411123,
+      "loss": 0.2532,
       "step": 835
     },
     {
-      "epoch": 10.501567398119121,
-      "grad_norm": 4.588846059050411e-05,
-      "learning_rate": 9.157350216771378e-05,
-      "loss": 0.0,
+      "epoch": 2.2395209580838324,
+      "grad_norm": 11.36906623840332,
+      "learning_rate": 0.00019480716290349995,
+      "loss": 0.3407,
       "step": 840
     },
     {
-      "epoch": 10.564263322884013,
-      "grad_norm": 0.0018681439105421305,
-      "learning_rate": 9.057317057921787e-05,
-      "loss": 0.0,
+      "epoch": 2.2528276779773786,
+      "grad_norm": 6.397517681121826,
+      "learning_rate": 0.00019473966425143292,
+      "loss": 0.2588,
       "step": 845
     },
     {
-      "epoch": 10.626959247648903,
-      "grad_norm": 0.0006441921577788889,
-      "learning_rate": 8.957378987628682e-05,
-      "loss": 0.0,
+      "epoch": 2.2661343978709247,
+      "grad_norm": 11.072936058044434,
+      "learning_rate": 0.00019467174158700504,
+      "loss": 0.2575,
       "step": 850
     },
     {
-      "epoch": 10.689655172413794,
-      "grad_norm": 0.0007152777980081737,
-      "learning_rate": 8.857546086658789e-05,
-      "loss": 0.0,
+      "epoch": 2.279441117764471,
+      "grad_norm": 0.4234752058982849,
+      "learning_rate": 0.00019460339521420772,
+      "loss": 0.1578,
       "step": 855
     },
     {
-      "epoch": 10.752351097178684,
-      "grad_norm": 0.0002644763153512031,
-      "learning_rate": 8.757828425170404e-05,
-      "loss": 0.0173,
+      "epoch": 2.2927478376580175,
+      "grad_norm": 52.92380905151367,
+      "learning_rate": 0.00019453462543892882,
+      "loss": 0.6933,
       "step": 860
     },
     {
-      "epoch": 10.815047021943574,
-      "grad_norm": 0.0004763313045259565,
-      "learning_rate": 8.658236061697586e-05,
-      "loss": 0.0,
+      "epoch": 2.3060545575515636,
+      "grad_norm": 18.455066680908203,
+      "learning_rate": 0.000194465432568951,
+      "loss": 0.3722,
       "step": 865
     },
     {
-      "epoch": 10.877742946708464,
-      "grad_norm": 0.07049086689949036,
-      "learning_rate": 8.55877904213558e-05,
-      "loss": 0.0,
+      "epoch": 2.31936127744511,
+      "grad_norm": 12.93152141571045,
+      "learning_rate": 0.00019439581691395067,
+      "loss": 0.3835,
       "step": 870
     },
     {
-      "epoch": 10.940438871473354,
-      "grad_norm": 0.00021191804262343794,
-      "learning_rate": 8.459467398727462e-05,
-      "loss": 0.0,
+      "epoch": 2.332667997338656,
+      "grad_norm": 5.5360212326049805,
+      "learning_rate": 0.00019432577878549637,
+      "loss": 0.339,
       "step": 875
     },
     {
-      "epoch": 11.0,
-      "grad_norm": 0.0024338788352906704,
-      "learning_rate": 8.360311149052205e-05,
-      "loss": 0.0,
-      "step": 880
-    },
-    {
-      "epoch": 11.0,
-      "eval_loss": 2.944465160369873,
-      "eval_macro_f1": 73.94965715163707,
-      "eval_macro_precision": 73.91418752481633,
-      "eval_macro_recall": 73.99999999999999,
-      "eval_micro_f1": 74.0,
-      "eval_micro_precision": 74.0,
-      "eval_micro_recall": 74.0,
-      "eval_runtime": 3.4955,
-      "eval_samples_per_second": 300.386,
-      "eval_steps_per_second": 18.881,
+      "epoch": 2.345974717232202,
+      "grad_norm": 6.381242752075195,
+      "learning_rate": 0.0001942553184970474,
+      "loss": 0.3283,
       "step": 880
     },
     {
-      "epoch": 11.06269592476489,
-      "grad_norm": 0.0007475401507690549,
-      "learning_rate": 8.261320295014182e-05,
-      "loss": 0.0,
+      "epoch": 2.3592814371257487,
+      "grad_norm": 7.571563720703125,
+      "learning_rate": 0.00019418443636395248,
+      "loss": 0.3261,
       "step": 885
     },
     {
-      "epoch": 11.12539184952978,
-      "grad_norm": 5.158692511031404e-05,
-      "learning_rate": 8.162504821834295e-05,
-      "loss": 0.0,
+      "epoch": 2.372588157019295,
+      "grad_norm": 2.1912384033203125,
+      "learning_rate": 0.00019411313270344837,
+      "loss": 0.3227,
       "step": 890
     },
     {
-      "epoch": 11.18808777429467,
-      "grad_norm": 0.0004298978892620653,
-      "learning_rate": 8.06387469704276e-05,
-      "loss": 0.0,
+      "epoch": 2.385894876912841,
+      "grad_norm": 4.941147804260254,
+      "learning_rate": 0.00019404140783465837,
+      "loss": 0.5089,
       "step": 895
     },
     {
-      "epoch": 11.25078369905956,
-      "grad_norm": 0.005391178652644157,
-      "learning_rate": 7.965439869473664e-05,
-      "loss": 0.0,
+      "epoch": 2.399201596806387,
+      "grad_norm": 6.276638507843018,
+      "learning_rate": 0.00019396926207859084,
+      "loss": 0.4068,
       "step": 900
     },
     {
-      "epoch": 11.31347962382445,
-      "grad_norm": 0.00016025469813030213,
-      "learning_rate": 7.867210268261439e-05,
-      "loss": 0.0,
+      "epoch": 2.4125083166999337,
+      "grad_norm": 4.92919397354126,
+      "learning_rate": 0.000193896695758138,
+      "loss": 0.4691,
       "step": 905
     },
     {
-      "epoch": 11.376175548589341,
-      "grad_norm": 0.00021810800535604358,
-      "learning_rate": 7.769195801839313e-05,
-      "loss": 0.0,
+      "epoch": 2.42581503659348,
+      "grad_norm": 6.015539169311523,
+      "learning_rate": 0.00019382370919807419,
+      "loss": 0.475,
       "step": 910
     },
     {
-      "epoch": 11.438871473354231,
-      "grad_norm": 0.0006552350823767483,
-      "learning_rate": 7.671406356939836e-05,
-      "loss": 0.0,
+      "epoch": 2.439121756487026,
+      "grad_norm": 16.059900283813477,
+      "learning_rate": 0.00019375030272505463,
+      "loss": 0.596,
       "step": 915
     },
     {
-      "epoch": 11.501567398119121,
-      "grad_norm": 0.001337722409516573,
-      "learning_rate": 7.573851797597602e-05,
-      "loss": 0.0,
+      "epoch": 2.452428476380572,
+      "grad_norm": 4.761904716491699,
+      "learning_rate": 0.00019367647666761385,
+      "loss": 0.6156,
       "step": 920
     },
     {
-      "epoch": 11.564263322884013,
-      "grad_norm": 0.00711465161293745,
-      "learning_rate": 7.476541964154269e-05,
-      "loss": 0.0,
+      "epoch": 2.4657351962741183,
+      "grad_norm": 11.512598037719727,
+      "learning_rate": 0.00019360223135616426,
+      "loss": 0.4051,
       "step": 925
     },
     {
-      "epoch": 11.626959247648903,
-      "grad_norm": 3.7509220419451594e-05,
-      "learning_rate": 7.379486672265964e-05,
-      "loss": 0.0,
+      "epoch": 2.479041916167665,
+      "grad_norm": 5.330412864685059,
+      "learning_rate": 0.00019352756712299468,
+      "loss": 0.2981,
       "step": 930
     },
     {
-      "epoch": 11.689655172413794,
-      "grad_norm": 0.00123989034909755,
-      "learning_rate": 7.28269571191317e-05,
-      "loss": 0.0,
+      "epoch": 2.492348636061211,
+      "grad_norm": 3.5500071048736572,
+      "learning_rate": 0.0001934524843022688,
+      "loss": 0.485,
       "step": 935
     },
     {
-      "epoch": 11.752351097178684,
-      "grad_norm": 0.00043695452040992677,
-      "learning_rate": 7.186178846413214e-05,
-      "loss": 0.0,
+      "epoch": 2.505655355954757,
+      "grad_norm": 6.358755111694336,
+      "learning_rate": 0.0001933769832300237,
+      "loss": 0.5665,
       "step": 940
     },
     {
-      "epoch": 11.815047021943574,
-      "grad_norm": 0.0020732777193188667,
-      "learning_rate": 7.089945811435433e-05,
-      "loss": 0.0,
+      "epoch": 2.5189620758483033,
+      "grad_norm": 12.199417114257812,
+      "learning_rate": 0.00019330106424416852,
+      "loss": 0.4757,
       "step": 945
     },
     {
-      "epoch": 11.877742946708464,
-      "grad_norm": 0.0006961887702345848,
-      "learning_rate": 6.994006314019141e-05,
-      "loss": 0.0,
+      "epoch": 2.5322687957418495,
+      "grad_norm": 11.047940254211426,
+      "learning_rate": 0.00019322472768448258,
+      "loss": 0.3742,
       "step": 950
     },
     {
-      "epoch": 11.940438871473354,
-      "grad_norm": 0.00014696276048198342,
-      "learning_rate": 6.898370031594487e-05,
-      "loss": 0.0,
+      "epoch": 2.545575515635396,
+      "grad_norm": 5.156216144561768,
+      "learning_rate": 0.00019314797389261424,
+      "loss": 0.4561,
       "step": 955
     },
     {
-      "epoch": 12.0,
-      "grad_norm": 0.004987122491002083,
-      "learning_rate": 6.803046611006278e-05,
-      "loss": 0.0,
-      "step": 960
-    },
-    {
-      "epoch": 12.0,
-      "eval_loss": 2.9617700576782227,
-      "eval_macro_f1": 74.48450982070439,
-      "eval_macro_precision": 74.49341191282451,
-      "eval_macro_recall": 74.47619047619047,
-      "eval_micro_f1": 74.47619047619047,
-      "eval_micro_precision": 74.47619047619047,
-      "eval_micro_recall": 74.47619047619047,
-      "eval_runtime": 3.4656,
-      "eval_samples_per_second": 302.974,
-      "eval_steps_per_second": 19.044,
+      "epoch": 2.5588822355289422,
+      "grad_norm": 5.482617378234863,
+      "learning_rate": 0.00019307080321207912,
+      "loss": 0.5187,
       "step": 960
     },
     {
-      "epoch": 12.06269592476489,
-      "grad_norm": 3.519210804370232e-05,
-      "learning_rate": 6.708045667540897e-05,
-      "loss": 0.0,
+      "epoch": 2.5721889554224884,
+      "grad_norm": 7.736295700073242,
+      "learning_rate": 0.00019299321598825866,
+      "loss": 0.4599,
       "step": 965
     },
     {
-      "epoch": 12.12539184952978,
-      "grad_norm": 0.0011250913375988603,
-      "learning_rate": 6.613376783956423e-05,
-      "loss": 0.0,
+      "epoch": 2.5854956753160345,
+      "grad_norm": 3.486142635345459,
+      "learning_rate": 0.00019291521256839858,
+      "loss": 0.2753,
       "step": 970
     },
     {
-      "epoch": 12.18808777429467,
-      "grad_norm": 5.886521466891281e-05,
-      "learning_rate": 6.519049509516013e-05,
-      "loss": 0.0,
+      "epoch": 2.5988023952095807,
+      "grad_norm": 4.562074184417725,
+      "learning_rate": 0.00019283679330160726,
+      "loss": 0.2733,
       "step": 975
     },
     {
-      "epoch": 12.25078369905956,
-      "grad_norm": 0.0004661143757402897,
-      "learning_rate": 6.425073359024663e-05,
-      "loss": 0.0,
+      "epoch": 2.6121091151031273,
+      "grad_norm": 13.783202171325684,
+      "learning_rate": 0.00019275795853885433,
+      "loss": 0.3718,
       "step": 980
     },
     {
-      "epoch": 12.31347962382445,
-      "grad_norm": 0.00029624096350744367,
-      "learning_rate": 6.331457811869437e-05,
-      "loss": 0.0,
+      "epoch": 2.6254158349966734,
+      "grad_norm": 9.088604927062988,
+      "learning_rate": 0.00019267870863296887,
+      "loss": 0.2797,
       "step": 985
     },
     {
-      "epoch": 12.376175548589341,
-      "grad_norm": 0.00012905469338875264,
-      "learning_rate": 6.2382123110633e-05,
-      "loss": 0.0,
+      "epoch": 2.6387225548902196,
+      "grad_norm": 10.746959686279297,
+      "learning_rate": 0.00019259904393863802,
+      "loss": 0.5173,
       "step": 990
     },
     {
-      "epoch": 12.438871473354231,
-      "grad_norm": 0.00013497307372745126,
-      "learning_rate": 6.145346262292595e-05,
-      "loss": 0.0,
+      "epoch": 2.6520292747837657,
+      "grad_norm": 29.163745880126953,
+      "learning_rate": 0.00019251896481240537,
+      "loss": 0.4255,
       "step": 995
     },
     {
-      "epoch": 12.501567398119121,
-      "grad_norm": 0.0008577414555475116,
-      "learning_rate": 6.052869032968285e-05,
-      "loss": 0.0,
+      "epoch": 2.665335994677312,
+      "grad_norm": 10.095052719116211,
+      "learning_rate": 0.0001924384716126692,
+      "loss": 0.6098,
       "step": 1000
     },
     {
-      "epoch": 12.564263322884013,
-      "grad_norm": 0.0001506131811765954,
-      "learning_rate": 5.960789951281052e-05,
-      "loss": 0.0,
+      "epoch": 2.6786427145708585,
+      "grad_norm": 4.459863185882568,
+      "learning_rate": 0.0001923575646996811,
+      "loss": 0.2524,
       "step": 1005
     },
     {
-      "epoch": 12.626959247648903,
-      "grad_norm": 0.00014009448932483792,
-      "learning_rate": 5.8691183052603834e-05,
-      "loss": 0.0,
+      "epoch": 2.6919494344644046,
+      "grad_norm": 8.43911361694336,
+      "learning_rate": 0.00019227624443554425,
+      "loss": 0.5336,
       "step": 1010
     },
     {
-      "epoch": 12.689655172413794,
-      "grad_norm": 2.588312781881541e-05,
-      "learning_rate": 5.777863341837675e-05,
-      "loss": 0.0,
+      "epoch": 2.7052561543579507,
+      "grad_norm": 7.737038612365723,
+      "learning_rate": 0.0001921945111842117,
+      "loss": 0.3688,
       "step": 1015
     },
     {
-      "epoch": 12.752351097178684,
-      "grad_norm": 0.00011443781841080636,
-      "learning_rate": 5.687034265913485e-05,
-      "loss": 0.0,
+      "epoch": 2.718562874251497,
+      "grad_norm": 10.94029712677002,
+      "learning_rate": 0.000192112365311485,
+      "loss": 0.4579,
       "step": 1020
     },
     {
-      "epoch": 12.815047021943574,
-      "grad_norm": 0.00010904014925472438,
-      "learning_rate": 5.596640239429051e-05,
-      "loss": 0.0,
+      "epoch": 2.731869594145043,
+      "grad_norm": 6.617719650268555,
+      "learning_rate": 0.0001920298071850123,
+      "loss": 0.1729,
       "step": 1025
     },
     {
-      "epoch": 12.877742946708464,
-      "grad_norm": 5.243328268988989e-05,
-      "learning_rate": 5.5066903804421025e-05,
-      "loss": 0.0,
+      "epoch": 2.7451763140385896,
+      "grad_norm": 15.398009300231934,
+      "learning_rate": 0.00019194683717428687,
+      "loss": 0.4216,
       "step": 1030
     },
     {
-      "epoch": 12.940438871473354,
-      "grad_norm": 0.0001976548373932019,
-      "learning_rate": 5.4171937622071435e-05,
-      "loss": 0.0,
+      "epoch": 2.758483033932136,
+      "grad_norm": 9.530508995056152,
+      "learning_rate": 0.00019186345565064535,
+      "loss": 0.5025,
       "step": 1035
     },
     {
-      "epoch": 13.0,
-      "grad_norm": 1.4458280929829925e-05,
-      "learning_rate": 5.32815941226022e-05,
-      "loss": 0.0,
-      "step": 1040
-    },
-    {
-      "epoch": 13.0,
-      "eval_loss": 2.9744179248809814,
-      "eval_macro_f1": 74.495166799055,
-      "eval_macro_precision": 74.51671347356447,
-      "eval_macro_recall": 74.47619047619047,
-      "eval_micro_f1": 74.47619047619047,
-      "eval_micro_precision": 74.47619047619047,
-      "eval_micro_recall": 74.47619047619047,
-      "eval_runtime": 3.7357,
-      "eval_samples_per_second": 281.069,
-      "eval_steps_per_second": 17.667,
+      "epoch": 2.771789753825682,
+      "grad_norm": 7.099113464355469,
+      "learning_rate": 0.00019177966298726613,
+      "loss": 0.3395,
       "step": 1040
     },
     {
-      "epoch": 13.06269592476489,
-      "grad_norm": 0.00015813493519090116,
-      "learning_rate": 5.2395963115083104e-05,
-      "loss": 0.0,
+      "epoch": 2.785096473719228,
+      "grad_norm": 15.377814292907715,
+      "learning_rate": 0.0001916954595591677,
+      "loss": 0.579,
       "step": 1045
     },
     {
-      "epoch": 13.12539184952978,
-      "grad_norm": 5.232647708908189e-06,
-      "learning_rate": 5.151513393323426e-05,
-      "loss": 0.0,
+      "epoch": 2.7984031936127742,
+      "grad_norm": 3.2470192909240723,
+      "learning_rate": 0.00019161084574320696,
+      "loss": 0.5129,
       "step": 1050
     },
     {
-      "epoch": 13.18808777429467,
-      "grad_norm": 0.0007055670721456409,
-      "learning_rate": 5.06391954264149e-05,
-      "loss": 0.0,
+      "epoch": 2.811709913506321,
+      "grad_norm": 8.268476486206055,
+      "learning_rate": 0.0001915258219180775,
+      "loss": 0.4809,
       "step": 1055
     },
     {
-      "epoch": 13.25078369905956,
-      "grad_norm": 0.0001715083053568378,
-      "learning_rate": 4.976823595066128e-05,
-      "loss": 0.0,
+      "epoch": 2.825016633399867,
+      "grad_norm": 23.91556739807129,
+      "learning_rate": 0.0001914403884643079,
+      "loss": 0.566,
       "step": 1060
     },
     {
-      "epoch": 13.31347962382445,
-      "grad_norm": 0.0010762359015643597,
-      "learning_rate": 4.8902343359774085e-05,
-      "loss": 0.0,
+      "epoch": 2.838323353293413,
+      "grad_norm": 5.273834228515625,
+      "learning_rate": 0.0001913545457642601,
+      "loss": 0.4376,
       "step": 1065
     },
     {
-      "epoch": 13.376175548589341,
-      "grad_norm": 0.0003584644291549921,
-      "learning_rate": 4.804160499645667e-05,
-      "loss": 0.0,
+      "epoch": 2.8516300731869593,
+      "grad_norm": 16.142292022705078,
+      "learning_rate": 0.00019126829420212764,
+      "loss": 0.2264,
       "step": 1070
     },
     {
-      "epoch": 13.438871473354231,
-      "grad_norm": 0.00013147966819815338,
-      "learning_rate": 4.7186107683504656e-05,
-      "loss": 0.0,
+      "epoch": 2.8649367930805054,
+      "grad_norm": 26.231626510620117,
+      "learning_rate": 0.00019118163416393392,
+      "loss": 0.3179,
       "step": 1075
     },
     {
-      "epoch": 13.501567398119121,
-      "grad_norm": 0.00026007898850366473,
-      "learning_rate": 4.6335937715048306e-05,
-      "loss": 0.0,
+      "epoch": 2.878243512974052,
+      "grad_norm": 6.147984981536865,
+      "learning_rate": 0.0001910945660375305,
+      "loss": 0.2849,
       "step": 1080
     },
     {
-      "epoch": 13.564263322884013,
-      "grad_norm": 8.957670797826722e-05,
-      "learning_rate": 4.549118084784788e-05,
-      "loss": 0.0,
+      "epoch": 2.891550232867598,
+      "grad_norm": 17.026046752929688,
+      "learning_rate": 0.0001910070902125954,
+      "loss": 0.5235,
       "step": 1085
     },
     {
-      "epoch": 13.626959247648903,
-      "grad_norm": 7.390981409116648e-06,
-      "learning_rate": 4.465192229264337e-05,
-      "loss": 0.0,
+      "epoch": 2.9048569527611443,
+      "grad_norm": 10.020215034484863,
+      "learning_rate": 0.0001909192070806313,
+      "loss": 0.3592,
       "step": 1090
     },
     {
-      "epoch": 13.689655172413794,
-      "grad_norm": 0.0010093646124005318,
-      "learning_rate": 4.381824670555934e-05,
-      "loss": 0.0,
+      "epoch": 2.9181636726546905,
+      "grad_norm": 12.995983123779297,
+      "learning_rate": 0.0001908309170349637,
+      "loss": 0.4643,
       "step": 1095
     },
     {
-      "epoch": 13.752351097178684,
-      "grad_norm": 0.0001321593881584704,
-      "learning_rate": 4.29902381795655e-05,
-      "loss": 0.0,
+      "epoch": 2.9314703925482366,
+      "grad_norm": 8.028570175170898,
+      "learning_rate": 0.00019074222047073947,
+      "loss": 0.6559,
       "step": 1100
     },
     {
-      "epoch": 13.815047021943574,
-      "grad_norm": 0.0003493933181744069,
-      "learning_rate": 4.216798023599441e-05,
-      "loss": 0.0,
+      "epoch": 2.944777112441783,
+      "grad_norm": 7.6127471923828125,
+      "learning_rate": 0.00019065311778492468,
+      "loss": 0.5201,
       "step": 1105
     },
     {
-      "epoch": 13.877742946708464,
-      "grad_norm": 0.00017224319162778556,
-      "learning_rate": 4.135155581611661e-05,
-      "loss": 0.0,
+      "epoch": 2.9580838323353293,
+      "grad_norm": 2.081876039505005,
+      "learning_rate": 0.0001905636093763031,
+      "loss": 0.2437,
       "step": 1110
     },
     {
-      "epoch": 13.940438871473354,
-      "grad_norm": 0.0007415884756483138,
-      "learning_rate": 4.0541047272774315e-05,
-      "loss": 0.0,
+      "epoch": 2.9713905522288755,
+      "grad_norm": 20.373952865600586,
+      "learning_rate": 0.00019047369564547436,
+      "loss": 0.8751,
       "step": 1115
     },
     {
-      "epoch": 14.0,
-      "grad_norm": 0.00015337667718995363,
-      "learning_rate": 3.973653636207437e-05,
-      "loss": 0.0,
+      "epoch": 2.984697272122422,
+      "grad_norm": 11.040226936340332,
+      "learning_rate": 0.00019038337699485208,
+      "loss": 0.2046,
       "step": 1120
     },
     {
-      "epoch": 14.0,
-      "eval_loss": 2.978665351867676,
-      "eval_macro_f1": 74.39737250761942,
-      "eval_macro_precision": 74.41561381963534,
-      "eval_macro_recall": 74.38095238095238,
-      "eval_micro_f1": 74.38095238095238,
-      "eval_micro_precision": 74.38095238095238,
-      "eval_micro_recall": 74.38095238095238,
-      "eval_runtime": 4.1748,
-      "eval_samples_per_second": 251.508,
-      "eval_steps_per_second": 15.809,
-      "step": 1120
+      "epoch": 2.998003992015968,
+      "grad_norm": 14.944093704223633,
+      "learning_rate": 0.00019029265382866214,
+      "loss": 0.5185,
+      "step": 1125
     },
     {
-      "epoch": 14.06269592476489,
-      "grad_norm": 0.00025141274090856314,
-      "learning_rate": 3.893810423514172e-05,
-      "loss": 0.0,
+      "epoch": 2.998003992015968,
+      "eval_loss": 0.556496798992157,
+      "eval_macro_f1": 67.9724393832533,
+      "eval_macro_precision": 66.8065205424496,
+      "eval_macro_recall": 69.37092559087907,
+      "eval_micro_f1": 86.00299401197606,
+      "eval_micro_precision": 86.00299401197606,
+      "eval_micro_recall": 86.00299401197606,
+      "eval_runtime": 4.5031,
+      "eval_samples_per_second": 593.365,
+      "eval_steps_per_second": 37.085,
       "step": 1125
     },
     {
-      "epoch": 14.12539184952978,
-      "grad_norm": 0.0002585667825769633,
-      "learning_rate": 3.814583142993352e-05,
-      "loss": 0.0,
+      "epoch": 3.013306719893546,
+      "grad_norm": 7.3535308837890625,
+      "learning_rate": 0.00019020152655294085,
+      "loss": 0.4967,
       "step": 1130
     },
     {
-      "epoch": 14.18808777429467,
-      "grad_norm": 1.575879105075728e-05,
-      "learning_rate": 3.7359797863115283e-05,
-      "loss": 0.0,
+      "epoch": 3.0266134397870923,
+      "grad_norm": 7.688624382019043,
+      "learning_rate": 0.00019010999557553317,
+      "loss": 0.2367,
       "step": 1135
     },
     {
-      "epoch": 14.25078369905956,
-      "grad_norm": 3.0083991077844985e-05,
-      "learning_rate": 3.6580082821999786e-05,
-      "loss": 0.0,
+      "epoch": 3.039920159680639,
+      "grad_norm": 7.046450138092041,
+      "learning_rate": 0.0001900180613060908,
+      "loss": 0.1927,
       "step": 1140
     },
     {
-      "epoch": 14.31347962382445,
-      "grad_norm": 0.0006605722010135651,
-      "learning_rate": 3.580676495654911e-05,
-      "loss": 0.0,
+      "epoch": 3.053226879574185,
+      "grad_norm": 3.205350637435913,
+      "learning_rate": 0.0001899257241560704,
+      "loss": 0.1934,
       "step": 1145
     },
     {
-      "epoch": 14.376175548589341,
-      "grad_norm": 0.00018690273282118142,
-      "learning_rate": 3.503992227144147e-05,
-      "loss": 0.0,
+      "epoch": 3.066533599467731,
+      "grad_norm": 17.599550247192383,
+      "learning_rate": 0.0001898329845387317,
+      "loss": 0.2969,
       "step": 1150
     },
     {
-      "epoch": 14.438871473354231,
-      "grad_norm": 0.00042844025301747024,
-      "learning_rate": 3.427963211820274e-05,
-      "loss": 0.0,
+      "epoch": 3.0798403193612773,
+      "grad_norm": 30.870227813720703,
+      "learning_rate": 0.00018973984286913584,
+      "loss": 0.3991,
       "step": 1155
     },
     {
-      "epoch": 14.501567398119121,
-      "grad_norm": 0.00033945144969038665,
-      "learning_rate": 3.352597118740404e-05,
-      "loss": 0.0,
+      "epoch": 3.0931470392548235,
+      "grad_norm": 1.632244348526001,
+      "learning_rate": 0.0001896462995641432,
+      "loss": 0.3399,
       "step": 1160
     },
     {
-      "epoch": 14.564263322884013,
-      "grad_norm": 5.347471233108081e-05,
-      "learning_rate": 3.277901550092581e-05,
-      "loss": 0.0,
+      "epoch": 3.10645375914837,
+      "grad_norm": 1.5939713716506958,
+      "learning_rate": 0.00018955235504241187,
+      "loss": 0.2284,
       "step": 1165
     },
     {
-      "epoch": 14.626959247648903,
-      "grad_norm": 0.0009476160048507154,
-      "learning_rate": 3.2038840404289705e-05,
-      "loss": 0.0,
+      "epoch": 3.1197604790419162,
+      "grad_norm": 5.369167327880859,
+      "learning_rate": 0.00018945800972439538,
+      "loss": 0.2094,
       "step": 1170
     },
     {
-      "epoch": 14.689655172413794,
-      "grad_norm": 0.0006543208146467805,
-      "learning_rate": 3.13055205590583e-05,
-      "loss": 0.0,
+      "epoch": 3.1330671989354624,
+      "grad_norm": 10.1832914352417,
+      "learning_rate": 0.00018936326403234125,
+      "loss": 0.1946,
       "step": 1175
     },
     {
-      "epoch": 14.752351097178684,
-      "grad_norm": 6.901170854689553e-05,
-      "learning_rate": 3.0579129935304066e-05,
-      "loss": 0.0,
+      "epoch": 3.1463739188290085,
+      "grad_norm": 0.6386124491691589,
+      "learning_rate": 0.00018926811839028876,
+      "loss": 0.1098,
       "step": 1180
     },
     {
-      "epoch": 14.815047021943574,
-      "grad_norm": 0.0006354337092489004,
-      "learning_rate": 2.9859741804147957e-05,
-      "loss": 0.0,
+      "epoch": 3.1596806387225547,
+      "grad_norm": 12.72846508026123,
+      "learning_rate": 0.00018917257322406734,
+      "loss": 0.1211,
       "step": 1185
     },
     {
-      "epoch": 14.877742946708464,
-      "grad_norm": 0.0009218246559612453,
-      "learning_rate": 2.9147428730368475e-05,
-      "loss": 0.0,
+      "epoch": 3.1729873586161013,
+      "grad_norm": 12.194605827331543,
+      "learning_rate": 0.00018907662896129433,
+      "loss": 0.4173,
       "step": 1190
     },
     {
-      "epoch": 14.940438871473354,
-      "grad_norm": 0.0008917151717469096,
-      "learning_rate": 2.844226256508221e-05,
-      "loss": 0.0,
+      "epoch": 3.1862940785096474,
+      "grad_norm": 4.248966217041016,
+      "learning_rate": 0.00018898028603137341,
+      "loss": 0.1053,
       "step": 1195
     },
     {
-      "epoch": 15.0,
-      "grad_norm": 4.064366657985374e-05,
-      "learning_rate": 2.7744314438496088e-05,
-      "loss": 0.0,
-      "step": 1200
-    },
-    {
-      "epoch": 15.0,
-      "eval_loss": 2.9775209426879883,
-      "eval_macro_f1": 74.59288643616037,
-      "eval_macro_precision": 74.61805414620287,
-      "eval_macro_recall": 74.57142857142857,
-      "eval_micro_f1": 74.57142857142857,
-      "eval_micro_precision": 74.57142857142857,
-      "eval_micro_recall": 74.57142857142857,
-      "eval_runtime": 4.2475,
-      "eval_samples_per_second": 247.204,
-      "eval_steps_per_second": 15.539,
+      "epoch": 3.1996007984031936,
+      "grad_norm": 3.8942484855651855,
+      "learning_rate": 0.00018888354486549237,
+      "loss": 0.1792,
       "step": 1200
     },
     {
-      "epoch": 15.06269592476489,
-      "grad_norm": 0.0005438003572635353,
-      "learning_rate": 2.7053654752732528e-05,
-      "loss": 0.0,
+      "epoch": 3.2129075182967397,
+      "grad_norm": 5.4043755531311035,
+      "learning_rate": 0.0001887864058966214,
+      "loss": 0.2308,
       "step": 1205
     },
     {
-      "epoch": 15.12539184952978,
-      "grad_norm": 6.86753774061799e-05,
-      "learning_rate": 2.6370353174727836e-05,
-      "loss": 0.0,
+      "epoch": 3.2262142381902863,
+      "grad_norm": 23.204421997070312,
+      "learning_rate": 0.00018868886955951115,
+      "loss": 0.2898,
       "step": 1210
     },
     {
-      "epoch": 15.18808777429467,
-      "grad_norm": 0.00032200937857851386,
-      "learning_rate": 2.5694478629205078e-05,
-      "loss": 0.0,
+      "epoch": 3.2395209580838324,
+      "grad_norm": 19.77983856201172,
+      "learning_rate": 0.00018859093629069058,
+      "loss": 0.1953,
       "step": 1215
     },
     {
-      "epoch": 15.25078369905956,
-      "grad_norm": 0.0002235895226476714,
-      "learning_rate": 2.5026099291721516e-05,
-      "loss": 0.0,
+      "epoch": 3.2528276779773786,
+      "grad_norm": 10.021352767944336,
+      "learning_rate": 0.00018849260652846519,
+      "loss": 0.3812,
       "step": 1220
     },
     {
-      "epoch": 15.31347962382445,
-      "grad_norm": 0.00034938243334181607,
-      "learning_rate": 2.4365282581791782e-05,
-      "loss": 0.0,
+      "epoch": 3.2661343978709247,
+      "grad_norm": 3.085806131362915,
+      "learning_rate": 0.00018839388071291503,
+      "loss": 0.1707,
       "step": 1225
     },
     {
-      "epoch": 15.376175548589341,
-      "grad_norm": 0.00023688429791945964,
-      "learning_rate": 2.371209515608718e-05,
-      "loss": 0.0,
+      "epoch": 3.279441117764471,
+      "grad_norm": 1.3998228311538696,
+      "learning_rate": 0.00018829475928589271,
+      "loss": 0.1516,
       "step": 1230
     },
     {
-      "epoch": 15.438871473354231,
-      "grad_norm": 0.0008163132588379085,
-      "learning_rate": 2.3066602901712108e-05,
-      "loss": 0.0,
+      "epoch": 3.2927478376580175,
+      "grad_norm": 6.1686835289001465,
+      "learning_rate": 0.00018819524269102136,
+      "loss": 0.3816,
       "step": 1235
     },
     {
-      "epoch": 15.501567398119121,
-      "grad_norm": 0.00019654417701531202,
-      "learning_rate": 2.242887092955801e-05,
-      "loss": 0.0,
+      "epoch": 3.3060545575515636,
+      "grad_norm": 6.946842670440674,
+      "learning_rate": 0.0001880953313736928,
+      "loss": 0.1497,
       "step": 1240
     },
     {
-      "epoch": 15.564263322884013,
-      "grad_norm": 2.644214873726014e-05,
-      "learning_rate": 2.1798963567735608e-05,
-      "loss": 0.0,
+      "epoch": 3.31936127744511,
+      "grad_norm": 3.2674038410186768,
+      "learning_rate": 0.00018799502578106534,
+      "loss": 0.1253,
       "step": 1245
     },
     {
-      "epoch": 15.626959247648903,
-      "grad_norm": 0.0005443547270260751,
-      "learning_rate": 2.1176944355086058e-05,
-      "loss": 0.0,
+      "epoch": 3.332667997338656,
+      "grad_norm": 4.695175647735596,
+      "learning_rate": 0.00018789432636206197,
+      "loss": 0.1789,
       "step": 1250
     },
     {
-      "epoch": 15.689655172413794,
-      "grad_norm": 0.00014729479153174907,
-      "learning_rate": 2.0562876034771882e-05,
-      "loss": 0.0,
+      "epoch": 3.345974717232202,
+      "grad_norm": 14.17712116241455,
+      "learning_rate": 0.00018779323356736826,
+      "loss": 0.1984,
       "step": 1255
     },
     {
-      "epoch": 15.752351097178684,
-      "grad_norm": 0.0004818035813514143,
-      "learning_rate": 1.995682054794803e-05,
-      "loss": 0.0,
+      "epoch": 3.3592814371257487,
+      "grad_norm": 9.122124671936035,
+      "learning_rate": 0.0001876917478494303,
+      "loss": 0.175,
       "step": 1260
     },
     {
-      "epoch": 15.815047021943574,
-      "grad_norm": 0.0002842153771780431,
-      "learning_rate": 1.935883902751382e-05,
-      "loss": 0.0,
+      "epoch": 3.372588157019295,
+      "grad_norm": 5.063784122467041,
+      "learning_rate": 0.00018758986966245283,
+      "loss": 0.2501,
       "step": 1265
     },
     {
-      "epoch": 15.877742946708464,
-      "grad_norm": 0.00016755808610469103,
-      "learning_rate": 1.8768991791946456e-05,
-      "loss": 0.0,
+      "epoch": 3.385894876912841,
+      "grad_norm": 8.168728828430176,
+      "learning_rate": 0.000187487599462397,
+      "loss": 0.2839,
       "step": 1270
     },
     {
-      "epoch": 15.940438871473354,
-      "grad_norm": 0.0001340256567345932,
-      "learning_rate": 1.8187338339216775e-05,
-      "loss": 0.0,
+      "epoch": 3.399201596806387,
+      "grad_norm": 6.399585247039795,
+      "learning_rate": 0.00018738493770697852,
+      "loss": 0.2172,
       "step": 1275
     },
     {
-      "epoch": 16.0,
-      "grad_norm": 0.00031726626912131906,
-      "learning_rate": 1.76139373407876e-05,
-      "loss": 0.0,
-      "step": 1280
-    },
-    {
-      "epoch": 16.0,
-      "eval_loss": 2.9834632873535156,
-      "eval_macro_f1": 74.30509548505665,
-      "eval_macro_precision": 74.3270405312326,
-      "eval_macro_recall": 74.28571428571429,
-      "eval_micro_f1": 74.28571428571429,
-      "eval_micro_precision": 74.28571428571429,
-      "eval_micro_recall": 74.28571428571429,
-      "eval_runtime": 5.3043,
-      "eval_samples_per_second": 197.954,
-      "eval_steps_per_second": 12.443,
+      "epoch": 3.4125083166999337,
+      "grad_norm": 3.3792059421539307,
+      "learning_rate": 0.00018728188485566544,
+      "loss": 0.3605,
       "step": 1280
     },
     {
-      "epoch": 16.062695924764892,
-      "grad_norm": 0.0002320996136404574,
-      "learning_rate": 1.7048846635695602e-05,
-      "loss": 0.0,
+      "epoch": 3.42581503659348,
+      "grad_norm": 2.7264702320098877,
+      "learning_rate": 0.00018717844136967624,
+      "loss": 0.3634,
       "step": 1285
     },
     {
-      "epoch": 16.12539184952978,
-      "grad_norm": 0.0005247556255199015,
-      "learning_rate": 1.649212322471695e-05,
-      "loss": 0.0,
+      "epoch": 3.439121756487026,
+      "grad_norm": 28.8140811920166,
+      "learning_rate": 0.00018707460771197774,
+      "loss": 0.2858,
       "step": 1290
     },
     {
-      "epoch": 16.188087774294672,
-      "grad_norm": 6.593632861040533e-05,
-      "learning_rate": 1.5943823264617796e-05,
-      "loss": 0.0,
+      "epoch": 3.452428476380572,
+      "grad_norm": 11.220648765563965,
+      "learning_rate": 0.0001869703843472829,
+      "loss": 0.2592,
       "step": 1295
     },
     {
-      "epoch": 16.25078369905956,
-      "grad_norm": 2.0959419998689555e-05,
-      "learning_rate": 1.540400206248963e-05,
-      "loss": 0.0,
+      "epoch": 3.4657351962741183,
+      "grad_norm": 6.689227104187012,
+      "learning_rate": 0.00018686577174204885,
+      "loss": 0.5982,
       "step": 1300
     },
     {
-      "epoch": 16.313479623824453,
-      "grad_norm": 0.0002277921448694542,
-      "learning_rate": 1.4872714070170468e-05,
-      "loss": 0.0,
+      "epoch": 3.479041916167665,
+      "grad_norm": 9.754223823547363,
+      "learning_rate": 0.00018676077036447494,
+      "loss": 0.1692,
       "step": 1305
     },
     {
-      "epoch": 16.37617554858934,
-      "grad_norm": 5.7650511735118926e-05,
-      "learning_rate": 1.435001287875234e-05,
-      "loss": 0.0,
+      "epoch": 3.492348636061211,
+      "grad_norm": 8.857412338256836,
+      "learning_rate": 0.00018665538068450023,
+      "loss": 0.6272,
       "step": 1310
     },
     {
-      "epoch": 16.438871473354233,
-      "grad_norm": 0.0003665912081487477,
-      "learning_rate": 1.3835951213175413e-05,
-      "loss": 0.0,
+      "epoch": 3.505655355954757,
+      "grad_norm": 11.387353897094727,
+      "learning_rate": 0.00018654960317380189,
+      "loss": 0.3136,
       "step": 1315
     },
     {
-      "epoch": 16.50156739811912,
-      "grad_norm": 0.00018977168656419963,
-      "learning_rate": 1.3330580926909763e-05,
-      "loss": 0.0,
+      "epoch": 3.5189620758483033,
+      "grad_norm": 12.20112419128418,
+      "learning_rate": 0.0001864434383057927,
+      "loss": 0.2563,
       "step": 1320
     },
     {
-      "epoch": 16.564263322884013,
-      "grad_norm": 2.3261072783498093e-05,
-      "learning_rate": 1.2833952996724863e-05,
-      "loss": 0.0,
+      "epoch": 3.5322687957418495,
+      "grad_norm": 5.7031121253967285,
+      "learning_rate": 0.0001863368865556191,
+      "loss": 0.2048,
       "step": 1325
     },
     {
-      "epoch": 16.6269592476489,
-      "grad_norm": 0.00015886298206169158,
-      "learning_rate": 1.2346117517547551e-05,
-      "loss": 0.0,
+      "epoch": 3.545575515635396,
+      "grad_norm": 36.38295364379883,
+      "learning_rate": 0.0001862299484001591,
+      "loss": 0.3182,
       "step": 1330
     },
     {
-      "epoch": 16.689655172413794,
-      "grad_norm": 0.00010217857925454155,
-      "learning_rate": 1.1867123697408854e-05,
-      "loss": 0.0,
+      "epoch": 3.5588822355289422,
+      "grad_norm": 0.3014439642429352,
+      "learning_rate": 0.00018612262431802007,
+      "loss": 0.1106,
       "step": 1335
     },
     {
-      "epoch": 16.752351097178682,
-      "grad_norm": 0.000697963812854141,
-      "learning_rate": 1.139701985248055e-05,
-      "loss": 0.0,
+      "epoch": 3.5721889554224884,
+      "grad_norm": 10.478448867797852,
+      "learning_rate": 0.00018601491478953657,
+      "loss": 0.3801,
       "step": 1340
     },
     {
-      "epoch": 16.815047021943574,
-      "grad_norm": 0.0002654240815900266,
-      "learning_rate": 1.0935853402201335e-05,
-      "loss": 0.0,
+      "epoch": 3.5854956753160345,
+      "grad_norm": 20.970182418823242,
+      "learning_rate": 0.00018590682029676823,
+      "loss": 0.2131,
       "step": 1345
     },
     {
-      "epoch": 16.877742946708462,
-      "grad_norm": 0.0005684046191163361,
-      "learning_rate": 1.0483670864493778e-05,
-      "loss": 0.0,
+      "epoch": 3.5988023952095807,
+      "grad_norm": 16.805198669433594,
+      "learning_rate": 0.00018579834132349772,
+      "loss": 0.4117,
       "step": 1350
     },
     {
-      "epoch": 16.940438871473354,
-      "grad_norm": 0.00042359635699540377,
-      "learning_rate": 1.004051785107184e-05,
-      "loss": 0.0,
+      "epoch": 3.6121091151031273,
+      "grad_norm": 6.421359539031982,
+      "learning_rate": 0.00018568947835522837,
+      "loss": 0.269,
       "step": 1355
     },
     {
-      "epoch": 17.0,
-      "grad_norm": 8.337834879057482e-05,
-      "learning_rate": 9.606439062840256e-06,
-      "loss": 0.0,
-      "step": 1360
-    },
-    {
-      "epoch": 17.0,
-      "eval_loss": 2.983590602874756,
-      "eval_macro_f1": 74.48693497595121,
-      "eval_macro_precision": 74.49848858038513,
-      "eval_macro_recall": 74.47619047619048,
-      "eval_micro_f1": 74.47619047619047,
-      "eval_micro_precision": 74.47619047619047,
-      "eval_micro_recall": 74.47619047619047,
-      "eval_runtime": 5.1696,
-      "eval_samples_per_second": 203.111,
-      "eval_steps_per_second": 12.767,
+      "epoch": 3.6254158349966734,
+      "grad_norm": 0.5675371885299683,
+      "learning_rate": 0.0001855802318791821,
+      "loss": 0.2104,
       "step": 1360
     },
     {
-      "epoch": 17.062695924764892,
-      "grad_norm": 0.0003075774875469506,
-      "learning_rate": 9.181478285385381e-06,
-      "loss": 0.0,
+      "epoch": 3.6387225548902196,
+      "grad_norm": 5.67156982421875,
+      "learning_rate": 0.00018547060238429736,
+      "loss": 0.248,
       "step": 1365
     },
     {
-      "epoch": 17.12539184952978,
-      "grad_norm": 8.384697139263153e-05,
-      "learning_rate": 8.765678384558607e-06,
-      "loss": 0.0,
+      "epoch": 3.6520292747837657,
+      "grad_norm": 5.464066982269287,
+      "learning_rate": 0.00018536059036122667,
+      "loss": 0.2855,
       "step": 1370
     },
     {
-      "epoch": 17.188087774294672,
-      "grad_norm": 0.00010231971828034148,
-      "learning_rate": 8.359081302152394e-06,
-      "loss": 0.0,
+      "epoch": 3.665335994677312,
+      "grad_norm": 2.300774097442627,
+      "learning_rate": 0.00018525019630233463,
+      "loss": 0.1769,
       "step": 1375
     },
     {
-      "epoch": 17.25078369905956,
-      "grad_norm": 0.0003627596015576273,
-      "learning_rate": 7.961728051669737e-06,
-      "loss": 0.0,
+      "epoch": 3.6786427145708585,
+      "grad_norm": 4.122674942016602,
+      "learning_rate": 0.0001851394207016957,
+      "loss": 0.3247,
       "step": 1380
     },
     {
-      "epoch": 17.313479623824453,
-      "grad_norm": 3.662167000584304e-05,
-      "learning_rate": 7.5736587141870155e-06,
-      "loss": 0.0,
+      "epoch": 3.6919494344644046,
+      "grad_norm": 4.85117244720459,
+      "learning_rate": 0.0001850282640550919,
+      "loss": 0.3428,
       "step": 1385
     },
     {
-      "epoch": 17.37617554858934,
-      "grad_norm": 0.0005837052594870329,
-      "learning_rate": 7.194912434311052e-06,
-      "loss": 0.0,
+      "epoch": 3.7052561543579507,
+      "grad_norm": 5.2849602699279785,
+      "learning_rate": 0.00018491672686001066,
+      "loss": 0.1914,
       "step": 1390
     },
     {
-      "epoch": 17.438871473354233,
-      "grad_norm": 7.577840005978942e-05,
-      "learning_rate": 6.8255274162305374e-06,
-      "loss": 0.0,
+      "epoch": 3.718562874251497,
+      "grad_norm": 4.865028381347656,
+      "learning_rate": 0.0001848048096156426,
+      "loss": 0.2503,
       "step": 1395
     },
     {
-      "epoch": 17.50156739811912,
-      "grad_norm": 0.00022036675363779068,
-      "learning_rate": 6.465540919862456e-06,
-      "loss": 0.0,
+      "epoch": 3.731869594145043,
+      "grad_norm": 5.3449482917785645,
+      "learning_rate": 0.0001846925128228792,
+      "loss": 0.2556,
       "step": 1400
     },
     {
-      "epoch": 17.564263322884013,
-      "grad_norm": 7.557481876574457e-05,
-      "learning_rate": 6.11498925709364e-06,
-      "loss": 0.0,
+      "epoch": 3.7451763140385896,
+      "grad_norm": 5.978095531463623,
+      "learning_rate": 0.00018457983698431075,
+      "loss": 0.0786,
       "step": 1405
     },
     {
-      "epoch": 17.6269592476489,
-      "grad_norm": 0.00023109870380721986,
-      "learning_rate": 5.77390778811796e-06,
-      "loss": 0.0,
+      "epoch": 3.758483033932136,
+      "grad_norm": 19.654855728149414,
+      "learning_rate": 0.00018446678260422385,
+      "loss": 0.4378,
       "step": 1410
     },
     {
-      "epoch": 17.689655172413794,
-      "grad_norm": 0.0002563658345025033,
-      "learning_rate": 5.44233091786951e-06,
-      "loss": 0.0,
+      "epoch": 3.771789753825682,
+      "grad_norm": 14.306594848632812,
+      "learning_rate": 0.0001843533501885993,
+      "loss": 0.4566,
       "step": 1415
     },
     {
-      "epoch": 17.752351097178682,
-      "grad_norm": 1.977133797481656e-05,
-      "learning_rate": 5.12029209255227e-06,
-      "loss": 0.0,
+      "epoch": 3.785096473719228,
+      "grad_norm": 17.783693313598633,
+      "learning_rate": 0.00018423954024510996,
+      "loss": 0.1464,
       "step": 1420
     },
     {
-      "epoch": 17.815047021943574,
-      "grad_norm": 0.00025862394249998033,
-      "learning_rate": 4.807823796266331e-06,
-      "loss": 0.0,
+      "epoch": 3.7984031936127742,
+      "grad_norm": 3.0436503887176514,
+      "learning_rate": 0.00018412535328311814,
+      "loss": 0.2691,
       "step": 1425
     },
     {
-      "epoch": 17.877742946708462,
-      "grad_norm": 0.00048169083311222494,
-      "learning_rate": 4.504957547731214e-06,
-      "loss": 0.0,
+      "epoch": 3.811709913506321,
+      "grad_norm": 18.622018814086914,
+      "learning_rate": 0.00018401078981367363,
+      "loss": 0.4671,
       "step": 1430
     },
     {
-      "epoch": 17.940438871473354,
-      "grad_norm": 1.255560255231103e-05,
-      "learning_rate": 4.211723897106534e-06,
-      "loss": 0.0,
+      "epoch": 3.825016633399867,
+      "grad_norm": 7.90670108795166,
+      "learning_rate": 0.0001838958503495113,
+      "loss": 0.3429,
       "step": 1435
     },
     {
-      "epoch": 18.0,
-      "grad_norm": 0.0004092851304449141,
-      "learning_rate": 3.928152422910491e-06,
-      "loss": 0.0,
-      "step": 1440
-    },
-    {
-      "epoch": 18.0,
-      "eval_loss": 2.982109308242798,
-      "eval_macro_f1": 74.49440308940996,
-      "eval_macro_precision": 74.5153120821081,
-      "eval_macro_recall": 74.47619047619047,
-      "eval_micro_f1": 74.47619047619047,
-      "eval_micro_precision": 74.47619047619047,
-      "eval_micro_recall": 74.47619047619047,
-      "eval_runtime": 5.0341,
-      "eval_samples_per_second": 208.578,
-      "eval_steps_per_second": 13.111,
+      "epoch": 3.838323353293413,
+      "grad_norm": 4.488195419311523,
+      "learning_rate": 0.00018378053540504873,
+      "loss": 0.3645,
       "step": 1440
     },
     {
-      "epoch": 18.062695924764892,
-      "grad_norm": 9.531569230603054e-05,
-      "learning_rate": 3.6542717290362515e-06,
-      "loss": 0.0,
+      "epoch": 3.8516300731869593,
+      "grad_norm": 8.05339241027832,
+      "learning_rate": 0.0001836648454963841,
+      "loss": 0.2885,
       "step": 1445
     },
     {
-      "epoch": 18.12539184952978,
-      "grad_norm": 5.50920121895615e-05,
-      "learning_rate": 3.390109441866618e-06,
-      "loss": 0.0,
+      "epoch": 3.8649367930805054,
+      "grad_norm": 3.8606925010681152,
+      "learning_rate": 0.00018354878114129367,
+      "loss": 0.2617,
       "step": 1450
     },
     {
-      "epoch": 18.188087774294672,
-      "grad_norm": 0.00010016823216574267,
-      "learning_rate": 3.135692207487373e-06,
-      "loss": 0.0,
+      "epoch": 3.878243512974052,
+      "grad_norm": 7.334080696105957,
+      "learning_rate": 0.00018343234285922953,
+      "loss": 0.2205,
       "step": 1455
     },
     {
-      "epoch": 18.25078369905956,
-      "grad_norm": 0.0006045085028745234,
-      "learning_rate": 2.8910456889995498e-06,
-      "loss": 0.0,
+      "epoch": 3.891550232867598,
+      "grad_norm": 10.274678230285645,
+      "learning_rate": 0.0001833155311713174,
+      "loss": 0.2675,
       "step": 1460
     },
     {
-      "epoch": 18.313479623824453,
-      "grad_norm": 0.00010283043229719624,
-      "learning_rate": 2.656194563930714e-06,
-      "loss": 0.0,
+      "epoch": 3.9048569527611443,
+      "grad_norm": 4.16168212890625,
+      "learning_rate": 0.00018319834660035413,
+      "loss": 0.1654,
       "step": 1465
     },
     {
-      "epoch": 18.37617554858934,
-      "grad_norm": 0.00047241951688192785,
-      "learning_rate": 2.4311625217457778e-06,
-      "loss": 0.0,
+      "epoch": 3.9181636726546905,
+      "grad_norm": 38.201255798339844,
+      "learning_rate": 0.00018308078967080546,
+      "loss": 0.2716,
       "step": 1470
     },
     {
-      "epoch": 18.438871473354233,
-      "grad_norm": 0.0003394366940483451,
-      "learning_rate": 2.2159722614573996e-06,
-      "loss": 0.0,
+      "epoch": 3.9314703925482366,
+      "grad_norm": 19.68052864074707,
+      "learning_rate": 0.0001829628609088036,
+      "loss": 0.4382,
       "step": 1475
     },
     {
-      "epoch": 18.50156739811912,
-      "grad_norm": 0.000773964449763298,
-      "learning_rate": 2.010645489336382e-06,
-      "loss": 0.0,
+      "epoch": 3.944777112441783,
+      "grad_norm": 4.916224956512451,
+      "learning_rate": 0.00018284456084214496,
+      "loss": 0.4168,
       "step": 1480
     },
     {
-      "epoch": 18.564263322884013,
-      "grad_norm": 6.701203528791666e-05,
-      "learning_rate": 1.8152029167221475e-06,
-      "loss": 0.0,
+      "epoch": 3.9580838323353293,
+      "grad_norm": 8.401510238647461,
+      "learning_rate": 0.00018272589000028772,
+      "loss": 0.4438,
       "step": 1485
     },
     {
-      "epoch": 18.6269592476489,
-      "grad_norm": 0.0003426434122957289,
-      "learning_rate": 1.6296642579335496e-06,
-      "loss": 0.0,
+      "epoch": 3.9713905522288755,
+      "grad_norm": 5.434607982635498,
+      "learning_rate": 0.0001826068489143495,
+      "loss": 0.3375,
       "step": 1490
     },
     {
-      "epoch": 18.689655172413794,
-      "grad_norm": 0.00010157287761103362,
-      "learning_rate": 1.4540482282803137e-06,
-      "loss": 0.0,
+      "epoch": 3.984697272122422,
+      "grad_norm": 2.88216495513916,
+      "learning_rate": 0.00018248743811710488,
+      "loss": 0.3334,
       "step": 1495
     },
     {
-      "epoch": 18.752351097178682,
-      "grad_norm": 0.0003385106392670423,
-      "learning_rate": 1.2883725421752201e-06,
-      "loss": 0.0,
+      "epoch": 3.998003992015968,
+      "grad_norm": 8.627459526062012,
+      "learning_rate": 0.0001823676581429833,
+      "loss": 0.3474,
       "step": 1500
     },
     {
-      "epoch": 18.815047021943574,
-      "grad_norm": 0.00031495324219577014,
-      "learning_rate": 1.132653911347248e-06,
-      "loss": 0.0,
+      "epoch": 3.998003992015968,
+      "eval_loss": 0.6241031289100647,
+      "eval_macro_f1": 65.75275685358942,
+      "eval_macro_precision": 66.95092371496796,
+      "eval_macro_recall": 66.23136102415846,
+      "eval_micro_f1": 85.55389221556887,
+      "eval_micro_precision": 85.55389221556887,
+      "eval_micro_recall": 85.55389221556887,
+      "eval_runtime": 4.7326,
+      "eval_samples_per_second": 564.595,
+      "eval_steps_per_second": 35.287,
+      "step": 1500
+    },
+    {
+      "epoch": 4.013306719893547,
+      "grad_norm": 1.1038553714752197,
+      "learning_rate": 0.00018224750952806624,
+      "loss": 0.0693,
       "step": 1505
     },
     {
-      "epoch": 18.877742946708462,
-      "grad_norm": 2.4275641408166848e-05,
-      "learning_rate": 9.869080431558542e-07,
-      "loss": 0.0,
+      "epoch": 4.026613439787092,
+      "grad_norm": 0.036140475422143936,
+      "learning_rate": 0.0001821269928100852,
+      "loss": 0.0629,
       "step": 1510
     },
     {
-      "epoch": 18.940438871473354,
-      "grad_norm": 0.0007264981977641582,
-      "learning_rate": 8.511496390065543e-07,
-      "loss": 0.0,
+      "epoch": 4.039920159680639,
+      "grad_norm": 16.097137451171875,
+      "learning_rate": 0.00018200610852841913,
+      "loss": 0.1647,
       "step": 1515
     },
     {
-      "epoch": 19.0,
-      "grad_norm": 3.2049887522589415e-05,
-      "learning_rate": 7.253923928680406e-07,
-      "loss": 0.0,
-      "step": 1520
-    },
-    {
-      "epoch": 19.0,
-      "eval_loss": 2.983579158782959,
-      "eval_macro_f1": 74.59288643616037,
-      "eval_macro_precision": 74.61805414620287,
-      "eval_macro_recall": 74.57142857142857,
-      "eval_micro_f1": 74.57142857142857,
-      "eval_micro_precision": 74.57142857142857,
-      "eval_micro_recall": 74.57142857142857,
-      "eval_runtime": 3.5338,
-      "eval_samples_per_second": 297.133,
-      "eval_steps_per_second": 18.677,
+      "epoch": 4.053226879574185,
+      "grad_norm": 97.72679138183594,
+      "learning_rate": 0.00018188485722409197,
+      "loss": 0.3568,
       "step": 1520
     },
     {
-      "epoch": 19.062695924764892,
-      "grad_norm": 0.00015775053179822862,
-      "learning_rate": 6.096489898908208e-07,
-      "loss": 0.0,
+      "epoch": 4.066533599467731,
+      "grad_norm": 26.921789169311523,
+      "learning_rate": 0.00018176323943977033,
+      "loss": 0.314,
       "step": 1525
     },
     {
-      "epoch": 19.12539184952978,
-      "grad_norm": 0.00020885077537968755,
-      "learning_rate": 5.039311051276752e-07,
-      "loss": 0.0,
+      "epoch": 4.079840319361278,
+      "grad_norm": 29.175439834594727,
+      "learning_rate": 0.00018164125571976098,
+      "loss": 0.151,
       "step": 1530
     },
     {
-      "epoch": 19.188087774294672,
-      "grad_norm": 0.0004631892079487443,
-      "learning_rate": 4.082494023560091e-07,
-      "loss": 0.0,
+      "epoch": 4.0931470392548235,
+      "grad_norm": 36.96811294555664,
+      "learning_rate": 0.00018151890661000856,
+      "loss": 0.1288,
       "step": 1535
     },
     {
-      "epoch": 19.25078369905956,
-      "grad_norm": 7.329711661441252e-05,
-      "learning_rate": 3.2261353300219176e-07,
-      "loss": 0.0,
+      "epoch": 4.10645375914837,
+      "grad_norm": 17.12455177307129,
+      "learning_rate": 0.0001813961926580929,
+      "loss": 0.1292,
       "step": 1540
     },
     {
-      "epoch": 19.313479623824453,
-      "grad_norm": 0.000103279686300084,
-      "learning_rate": 2.4703213516799053e-07,
-      "loss": 0.0,
+      "epoch": 4.119760479041916,
+      "grad_norm": 6.079716205596924,
+      "learning_rate": 0.0001812731144132268,
+      "loss": 0.2966,
       "step": 1545
     },
     {
-      "epoch": 19.37617554858934,
-      "grad_norm": 0.00035780860343948007,
-      "learning_rate": 1.8151283275928964e-07,
-      "loss": 0.0,
+      "epoch": 4.133067198935462,
+      "grad_norm": 6.217795372009277,
+      "learning_rate": 0.00018114967242625343,
+      "loss": 0.0419,
       "step": 1550
     },
     {
-      "epoch": 19.438871473354233,
-      "grad_norm": 0.0009104281198233366,
-      "learning_rate": 1.2606223471702817e-07,
-      "loss": 0.0,
+      "epoch": 4.146373918829009,
+      "grad_norm": 4.027215957641602,
+      "learning_rate": 0.00018102586724964387,
+      "loss": 0.1433,
       "step": 1555
     },
     {
-      "epoch": 19.50156739811912,
-      "grad_norm": 0.0008115767268463969,
-      "learning_rate": 8.068593435055505e-08,
-      "loss": 0.0,
+      "epoch": 4.159680638722555,
+      "grad_norm": 0.21189923584461212,
+      "learning_rate": 0.00018090169943749476,
+      "loss": 0.2901,
       "step": 1560
     },
     {
-      "epoch": 19.564263322884013,
-      "grad_norm": 2.108391527144704e-05,
-      "learning_rate": 4.5388508773469564e-08,
-      "loss": 0.0,
+      "epoch": 4.172987358616101,
+      "grad_norm": 0.8172288537025452,
+      "learning_rate": 0.00018077716954552564,
+      "loss": 0.2699,
       "step": 1565
     },
     {
-      "epoch": 19.6269592476489,
-      "grad_norm": 0.00026582309510558844,
-      "learning_rate": 2.0173518441868324e-08,
-      "loss": 0.0,
+      "epoch": 4.186294078509647,
+      "grad_norm": 1.5100902318954468,
+      "learning_rate": 0.00018065227813107666,
+      "loss": 0.1433,
       "step": 1570
     },
     {
-      "epoch": 19.689655172413794,
-      "grad_norm": 0.0001829984103096649,
-      "learning_rate": 5.043506795276987e-09,
-      "loss": 0.0,
+      "epoch": 4.199600798403194,
+      "grad_norm": 19.998355865478516,
+      "learning_rate": 0.00018052702575310588,
+      "loss": 0.2386,
       "step": 1575
     },
     {
-      "epoch": 19.752351097178682,
-      "grad_norm": 0.00031997705809772015,
-      "learning_rate": 0.0,
-      "loss": 0.0,
+      "epoch": 4.21290751829674,
+      "grad_norm": 2.6985023021698,
+      "learning_rate": 0.00018040141297218695,
+      "loss": 0.0853,
       "step": 1580
     },
     {
-      "epoch": 19.752351097178682,
-      "eval_loss": 2.983431816101074,
-      "eval_macro_f1": 74.495166799055,
-      "eval_macro_precision": 74.51671347356447,
-      "eval_macro_recall": 74.47619047619047,
-      "eval_micro_f1": 74.47619047619047,
-      "eval_micro_precision": 74.47619047619047,
-      "eval_micro_recall": 74.47619047619047,
-      "eval_runtime": 5.0918,
-      "eval_samples_per_second": 206.215,
-      "eval_steps_per_second": 12.962,
-      "step": 1580
+      "epoch": 4.226214238190286,
+      "grad_norm": 3.395392894744873,
+      "learning_rate": 0.00018027544035050644,
+      "loss": 0.1964,
+      "step": 1585
+    },
+    {
+      "epoch": 4.2395209580838324,
+      "grad_norm": 7.006717681884766,
+      "learning_rate": 0.00018014910845186153,
+      "loss": 0.2763,
+      "step": 1590
+    },
+    {
+      "epoch": 4.252827677977379,
+      "grad_norm": 1.993622899055481,
+      "learning_rate": 0.00018002241784165728,
+      "loss": 0.2098,
+      "step": 1595
+    },
+    {
+      "epoch": 4.266134397870925,
+      "grad_norm": 0.10528213530778885,
+      "learning_rate": 0.00017989536908690412,
+      "loss": 0.2781,
+      "step": 1600
+    },
+    {
+      "epoch": 4.279441117764471,
+      "grad_norm": 1.242319107055664,
+      "learning_rate": 0.00017976796275621555,
+      "loss": 0.029,
+      "step": 1605
+    },
+    {
+      "epoch": 4.292747837658017,
+      "grad_norm": 6.547476291656494,
+      "learning_rate": 0.00017964019941980522,
+      "loss": 0.4249,
+      "step": 1610
+    },
+    {
+      "epoch": 4.306054557551564,
+      "grad_norm": 6.174795150756836,
+      "learning_rate": 0.0001795120796494848,
+      "loss": 0.2151,
+      "step": 1615
+    },
+    {
+      "epoch": 4.319361277445109,
+      "grad_norm": 34.997474670410156,
+      "learning_rate": 0.00017938360401866093,
+      "loss": 0.0941,
+      "step": 1620
+    },
+    {
+      "epoch": 4.332667997338656,
+      "grad_norm": 4.191906452178955,
+      "learning_rate": 0.00017925477310233316,
+      "loss": 0.0643,
+      "step": 1625
+    },
+    {
+      "epoch": 4.3459747172322025,
+      "grad_norm": 0.43293797969818115,
+      "learning_rate": 0.00017912558747709104,
+      "loss": 0.1264,
+      "step": 1630
+    },
+    {
+      "epoch": 4.359281437125748,
+      "grad_norm": 24.775293350219727,
+      "learning_rate": 0.00017899604772111163,
+      "loss": 0.3707,
+      "step": 1635
+    },
+    {
+      "epoch": 4.372588157019295,
+      "grad_norm": 1.0308302640914917,
+      "learning_rate": 0.0001788661544141569,
+      "loss": 0.4379,
+      "step": 1640
+    },
+    {
+      "epoch": 4.385894876912841,
+      "grad_norm": 7.142487525939941,
+      "learning_rate": 0.0001787359081375713,
+      "loss": 0.3458,
+      "step": 1645
+    },
+    {
+      "epoch": 4.399201596806387,
+      "grad_norm": 18.565332412719727,
+      "learning_rate": 0.00017860530947427875,
+      "loss": 0.241,
+      "step": 1650
+    },
+    {
+      "epoch": 4.412508316699934,
+      "grad_norm": 1.515366554260254,
+      "learning_rate": 0.00017847435900878058,
+      "loss": 0.2142,
+      "step": 1655
+    },
+    {
+      "epoch": 4.425815036593479,
+      "grad_norm": 2.90205454826355,
+      "learning_rate": 0.0001783430573271524,
+      "loss": 0.1881,
+      "step": 1660
+    },
+    {
+      "epoch": 4.439121756487026,
+      "grad_norm": 3.373195171356201,
+      "learning_rate": 0.00017821140501704194,
+      "loss": 0.2427,
+      "step": 1665
+    },
+    {
+      "epoch": 4.452428476380573,
+      "grad_norm": 4.514166831970215,
+      "learning_rate": 0.00017807940266766593,
+      "loss": 0.2774,
+      "step": 1670
+    },
+    {
+      "epoch": 4.465735196274118,
+      "grad_norm": 16.700340270996094,
+      "learning_rate": 0.0001779470508698079,
+      "loss": 0.1988,
+      "step": 1675
+    },
+    {
+      "epoch": 4.479041916167665,
+      "grad_norm": 7.546816349029541,
+      "learning_rate": 0.00017781435021581527,
+      "loss": 0.2984,
+      "step": 1680
+    },
+    {
+      "epoch": 4.492348636061211,
+      "grad_norm": 1.9645479917526245,
+      "learning_rate": 0.00017768130129959683,
+      "loss": 0.3073,
+      "step": 1685
+    },
+    {
+      "epoch": 4.505655355954757,
+      "grad_norm": 37.8007926940918,
+      "learning_rate": 0.00017754790471662002,
+      "loss": 0.3818,
+      "step": 1690
+    },
+    {
+      "epoch": 4.518962075848304,
+      "grad_norm": 8.654253959655762,
+      "learning_rate": 0.00017741416106390826,
+      "loss": 0.5542,
+      "step": 1695
+    },
+    {
+      "epoch": 4.5322687957418495,
+      "grad_norm": 7.000682830810547,
+      "learning_rate": 0.0001772800709400383,
+      "loss": 0.5024,
+      "step": 1700
+    },
+    {
+      "epoch": 4.545575515635396,
+      "grad_norm": 11.131951332092285,
+      "learning_rate": 0.00017714563494513764,
+      "loss": 0.2344,
+      "step": 1705
+    },
+    {
+      "epoch": 4.558882235528942,
+      "grad_norm": 4.567352771759033,
+      "learning_rate": 0.00017701085368088156,
+      "loss": 0.2835,
+      "step": 1710
+    },
+    {
+      "epoch": 4.572188955422488,
+      "grad_norm": 12.048651695251465,
+      "learning_rate": 0.0001768757277504908,
+      "loss": 0.1937,
+      "step": 1715
+    },
+    {
+      "epoch": 4.585495675316035,
+      "grad_norm": 21.26102638244629,
+      "learning_rate": 0.00017674025775872852,
+      "loss": 0.3544,
+      "step": 1720
+    },
+    {
+      "epoch": 4.598802395209581,
+      "grad_norm": 8.841492652893066,
+      "learning_rate": 0.0001766044443118978,
+      "loss": 0.3688,
+      "step": 1725
+    },
+    {
+      "epoch": 4.612109115103127,
+      "grad_norm": 7.04481315612793,
+      "learning_rate": 0.00017646828801783895,
+      "loss": 0.2074,
+      "step": 1730
+    },
+    {
+      "epoch": 4.625415834996673,
+      "grad_norm": 6.048544406890869,
+      "learning_rate": 0.0001763317894859266,
+      "loss": 0.1041,
+      "step": 1735
+    },
+    {
+      "epoch": 4.63872255489022,
+      "grad_norm": 15.356284141540527,
+      "learning_rate": 0.0001761949493270671,
+      "loss": 0.1728,
+      "step": 1740
+    },
+    {
+      "epoch": 4.652029274783766,
+      "grad_norm": 1.0128647089004517,
+      "learning_rate": 0.00017605776815369581,
+      "loss": 0.1081,
+      "step": 1745
+    },
+    {
+      "epoch": 4.665335994677312,
+      "grad_norm": 23.05744743347168,
+      "learning_rate": 0.00017592024657977432,
+      "loss": 0.2708,
+      "step": 1750
+    },
+    {
+      "epoch": 4.6786427145708585,
+      "grad_norm": 17.319061279296875,
+      "learning_rate": 0.0001757823852207877,
+      "loss": 0.1197,
+      "step": 1755
+    },
+    {
+      "epoch": 4.691949434464404,
+      "grad_norm": 6.973044395446777,
+      "learning_rate": 0.00017564418469374167,
+      "loss": 0.4926,
+      "step": 1760
+    },
+    {
+      "epoch": 4.705256154357951,
+      "grad_norm": 44.00906753540039,
+      "learning_rate": 0.00017550564561716,
+      "loss": 0.1962,
+      "step": 1765
+    },
+    {
+      "epoch": 4.718562874251497,
+      "grad_norm": 3.3621838092803955,
+      "learning_rate": 0.00017536676861108164,
+      "loss": 0.181,
+      "step": 1770
+    },
+    {
+      "epoch": 4.731869594145043,
+      "grad_norm": 8.182963371276855,
+      "learning_rate": 0.00017522755429705798,
+      "loss": 0.0915,
+      "step": 1775
+    },
+    {
+      "epoch": 4.74517631403859,
+      "grad_norm": 27.33730697631836,
+      "learning_rate": 0.00017508800329814995,
+      "loss": 0.3047,
+      "step": 1780
+    },
+    {
+      "epoch": 4.758483033932135,
+      "grad_norm": 20.636539459228516,
+      "learning_rate": 0.0001749481162389254,
+      "loss": 0.3657,
+      "step": 1785
+    },
+    {
+      "epoch": 4.771789753825682,
+      "grad_norm": 26.856287002563477,
+      "learning_rate": 0.00017480789374545633,
+      "loss": 0.4397,
+      "step": 1790
+    },
+    {
+      "epoch": 4.7850964737192285,
+      "grad_norm": 6.763969898223877,
+      "learning_rate": 0.0001746673364453158,
+      "loss": 0.2756,
+      "step": 1795
+    },
+    {
+      "epoch": 4.798403193612774,
+      "grad_norm": 2.868919849395752,
+      "learning_rate": 0.0001745264449675755,
+      "loss": 0.1183,
+      "step": 1800
+    },
+    {
+      "epoch": 4.811709913506321,
+      "grad_norm": 13.854222297668457,
+      "learning_rate": 0.00017438521994280257,
+      "loss": 0.2545,
+      "step": 1805
+    },
+    {
+      "epoch": 4.825016633399867,
+      "grad_norm": 14.696944236755371,
+      "learning_rate": 0.00017424366200305714,
+      "loss": 0.3584,
+      "step": 1810
+    },
+    {
+      "epoch": 4.838323353293413,
+      "grad_norm": 18.185508728027344,
+      "learning_rate": 0.00017410177178188918,
+      "loss": 0.4049,
+      "step": 1815
+    },
+    {
+      "epoch": 4.85163007318696,
+      "grad_norm": 3.704495429992676,
+      "learning_rate": 0.00017395954991433586,
+      "loss": 0.3104,
+      "step": 1820
+    },
+    {
+      "epoch": 4.864936793080505,
+      "grad_norm": 16.54859161376953,
+      "learning_rate": 0.00017381699703691866,
+      "loss": 0.3405,
+      "step": 1825
+    },
+    {
+      "epoch": 4.878243512974052,
+      "grad_norm": 8.207486152648926,
+      "learning_rate": 0.0001736741137876405,
+      "loss": 0.2789,
+      "step": 1830
+    },
+    {
+      "epoch": 4.891550232867598,
+      "grad_norm": 2.0907087326049805,
+      "learning_rate": 0.0001735309008059829,
+      "loss": 0.1476,
+      "step": 1835
+    },
+    {
+      "epoch": 4.904856952761144,
+      "grad_norm": 7.623042106628418,
+      "learning_rate": 0.00017338735873290303,
+      "loss": 0.3736,
+      "step": 1840
+    },
+    {
+      "epoch": 4.918163672654691,
+      "grad_norm": 6.89530086517334,
+      "learning_rate": 0.0001732434882108311,
+      "loss": 0.1917,
+      "step": 1845
+    },
+    {
+      "epoch": 4.931470392548237,
+      "grad_norm": 46.501251220703125,
+      "learning_rate": 0.00017309928988366716,
+      "loss": 0.1845,
+      "step": 1850
+    },
+    {
+      "epoch": 4.944777112441783,
+      "grad_norm": 14.652153015136719,
+      "learning_rate": 0.00017295476439677847,
+      "loss": 0.2152,
+      "step": 1855
+    },
+    {
+      "epoch": 4.95808383233533,
+      "grad_norm": 16.384225845336914,
+      "learning_rate": 0.00017280991239699642,
+      "loss": 0.3643,
+      "step": 1860
+    },
+    {
+      "epoch": 4.9713905522288755,
+      "grad_norm": 1.2959167957305908,
+      "learning_rate": 0.00017266473453261377,
+      "loss": 0.341,
+      "step": 1865
+    },
+    {
+      "epoch": 4.984697272122422,
+      "grad_norm": 16.964174270629883,
+      "learning_rate": 0.00017251923145338176,
+      "loss": 0.393,
+      "step": 1870
+    },
+    {
+      "epoch": 4.998003992015968,
+      "grad_norm": 3.103013038635254,
+      "learning_rate": 0.00017237340381050703,
+      "loss": 0.2354,
+      "step": 1875
+    },
+    {
+      "epoch": 4.998003992015968,
+      "eval_loss": 0.6098228693008423,
+      "eval_macro_f1": 64.20293730741194,
+      "eval_macro_precision": 67.9233057601082,
+      "eval_macro_recall": 61.45233240477489,
+      "eval_micro_f1": 86.78892215568862,
+      "eval_micro_precision": 86.78892215568862,
+      "eval_micro_recall": 86.78892215568862,
+      "eval_runtime": 4.8204,
+      "eval_samples_per_second": 554.308,
+      "eval_steps_per_second": 34.644,
+      "step": 1875
+    },
+    {
+      "epoch": 5.013306719893547,
+      "grad_norm": 10.461470603942871,
+      "learning_rate": 0.0001722272522566489,
+      "loss": 0.1565,
+      "step": 1880
+    },
+    {
+      "epoch": 5.026613439787092,
+      "grad_norm": 3.9339942932128906,
+      "learning_rate": 0.00017208077744591634,
+      "loss": 0.1466,
+      "step": 1885
+    },
+    {
+      "epoch": 5.039920159680639,
+      "grad_norm": 7.42096471786499,
+      "learning_rate": 0.0001719339800338651,
+      "loss": 0.0382,
+      "step": 1890
+    },
+    {
+      "epoch": 5.053226879574185,
+      "grad_norm": 0.58461594581604,
+      "learning_rate": 0.0001717868606774948,
+      "loss": 0.1592,
+      "step": 1895
+    },
+    {
+      "epoch": 5.066533599467731,
+      "grad_norm": 1.377388596534729,
+      "learning_rate": 0.00017163942003524572,
+      "loss": 0.0645,
+      "step": 1900
+    },
+    {
+      "epoch": 5.079840319361278,
+      "grad_norm": 0.42152106761932373,
+      "learning_rate": 0.00017149165876699635,
+      "loss": 0.0022,
+      "step": 1905
+    },
+    {
+      "epoch": 5.0931470392548235,
+      "grad_norm": 0.027353579178452492,
+      "learning_rate": 0.00017134357753406003,
+      "loss": 0.2095,
+      "step": 1910
+    },
+    {
+      "epoch": 5.10645375914837,
+      "grad_norm": 44.416969299316406,
+      "learning_rate": 0.00017119517699918207,
+      "loss": 0.0876,
+      "step": 1915
+    },
+    {
+      "epoch": 5.119760479041916,
+      "grad_norm": 19.71007537841797,
+      "learning_rate": 0.0001710464578265369,
+      "loss": 0.0498,
+      "step": 1920
+    },
+    {
+      "epoch": 5.133067198935462,
+      "grad_norm": 8.512078285217285,
+      "learning_rate": 0.000170897420681725,
+      "loss": 0.2547,
+      "step": 1925
+    },
+    {
+      "epoch": 5.146373918829009,
+      "grad_norm": 0.14262406527996063,
+      "learning_rate": 0.00017074806623177,
+      "loss": 0.5994,
+      "step": 1930
+    },
+    {
+      "epoch": 5.159680638722555,
+      "grad_norm": 3.3490352630615234,
+      "learning_rate": 0.00017059839514511565,
+      "loss": 0.2067,
+      "step": 1935
+    },
+    {
+      "epoch": 5.172987358616101,
+      "grad_norm": 13.887410163879395,
+      "learning_rate": 0.00017044840809162271,
+      "loss": 0.3215,
+      "step": 1940
+    },
+    {
+      "epoch": 5.186294078509647,
+      "grad_norm": 1.4734302759170532,
+      "learning_rate": 0.0001702981057425662,
+      "loss": 0.1817,
+      "step": 1945
+    },
+    {
+      "epoch": 5.199600798403194,
+      "grad_norm": 30.033546447753906,
+      "learning_rate": 0.00017014748877063214,
+      "loss": 0.1956,
+      "step": 1950
+    },
+    {
+      "epoch": 5.21290751829674,
+      "grad_norm": 15.2096529006958,
+      "learning_rate": 0.00016999655784991478,
+      "loss": 0.2823,
+      "step": 1955
+    },
+    {
+      "epoch": 5.226214238190286,
+      "grad_norm": 40.786888122558594,
+      "learning_rate": 0.0001698453136559134,
+      "loss": 0.3874,
+      "step": 1960
+    },
+    {
+      "epoch": 5.2395209580838324,
+      "grad_norm": 21.138729095458984,
+      "learning_rate": 0.00016969375686552937,
+      "loss": 0.1977,
+      "step": 1965
+    },
+    {
+      "epoch": 5.252827677977379,
+      "grad_norm": 39.89869689941406,
+      "learning_rate": 0.00016954188815706305,
+      "loss": 0.2877,
+      "step": 1970
+    },
+    {
+      "epoch": 5.266134397870925,
+      "grad_norm": 4.198141098022461,
+      "learning_rate": 0.0001693897082102109,
+      "loss": 0.1792,
+      "step": 1975
+    },
+    {
+      "epoch": 5.279441117764471,
+      "grad_norm": 25.046525955200195,
+      "learning_rate": 0.00016923721770606228,
+      "loss": 0.1408,
+      "step": 1980
+    },
+    {
+      "epoch": 5.292747837658017,
+      "grad_norm": 31.608402252197266,
+      "learning_rate": 0.0001690844173270965,
+      "loss": 0.2465,
+      "step": 1985
+    },
+    {
+      "epoch": 5.306054557551564,
+      "grad_norm": 0.058475542813539505,
+      "learning_rate": 0.00016893130775717962,
+      "loss": 0.1367,
+      "step": 1990
+    },
+    {
+      "epoch": 5.319361277445109,
+      "grad_norm": 23.879507064819336,
+      "learning_rate": 0.0001687778896815617,
+      "loss": 0.1272,
+      "step": 1995
+    },
+    {
+      "epoch": 5.332667997338656,
+      "grad_norm": 11.176948547363281,
+      "learning_rate": 0.0001686241637868734,
+      "loss": 0.2328,
+      "step": 2000
+    },
+    {
+      "epoch": 5.3459747172322025,
+      "grad_norm": 62.34965515136719,
+      "learning_rate": 0.000168470130761123,
+      "loss": 0.1621,
+      "step": 2005
+    },
+    {
+      "epoch": 5.359281437125748,
+      "grad_norm": 18.16000747680664,
+      "learning_rate": 0.00016831579129369346,
+      "loss": 0.2798,
+      "step": 2010
+    },
+    {
+      "epoch": 5.372588157019295,
+      "grad_norm": 30.744770050048828,
+      "learning_rate": 0.00016816114607533922,
+      "loss": 0.2956,
+      "step": 2015
+    },
+    {
+      "epoch": 5.385894876912841,
+      "grad_norm": 2.792100429534912,
+      "learning_rate": 0.00016800619579818312,
+      "loss": 0.1253,
+      "step": 2020
+    },
+    {
+      "epoch": 5.399201596806387,
+      "grad_norm": 0.24668025970458984,
+      "learning_rate": 0.00016785094115571322,
+      "loss": 0.1933,
+      "step": 2025
+    },
+    {
+      "epoch": 5.412508316699934,
+      "grad_norm": 9.7402925491333,
+      "learning_rate": 0.00016769538284277995,
+      "loss": 0.0581,
+      "step": 2030
+    },
+    {
+      "epoch": 5.425815036593479,
+      "grad_norm": 1.663047432899475,
+      "learning_rate": 0.00016753952155559266,
+      "loss": 0.2041,
+      "step": 2035
+    },
+    {
+      "epoch": 5.439121756487026,
+      "grad_norm": 3.0766444206237793,
+      "learning_rate": 0.00016738335799171682,
+      "loss": 0.4174,
+      "step": 2040
+    },
+    {
+      "epoch": 5.452428476380573,
+      "grad_norm": 50.167137145996094,
+      "learning_rate": 0.0001672268928500706,
+      "loss": 0.2974,
+      "step": 2045
+    },
+    {
+      "epoch": 5.465735196274118,
+      "grad_norm": 1.0232924222946167,
+      "learning_rate": 0.00016707012683092208,
+      "loss": 0.2376,
+      "step": 2050
+    },
+    {
+      "epoch": 5.479041916167665,
+      "grad_norm": 5.915070056915283,
+      "learning_rate": 0.00016691306063588583,
+      "loss": 0.0967,
+      "step": 2055
+    },
+    {
+      "epoch": 5.492348636061211,
+      "grad_norm": 3.8537585735321045,
+      "learning_rate": 0.00016675569496791984,
+      "loss": 0.3152,
+      "step": 2060
+    },
+    {
+      "epoch": 5.505655355954757,
+      "grad_norm": 1.8343299627304077,
+      "learning_rate": 0.00016659803053132249,
+      "loss": 0.1102,
+      "step": 2065
+    },
+    {
+      "epoch": 5.518962075848304,
+      "grad_norm": 3.2312400341033936,
+      "learning_rate": 0.00016644006803172924,
+      "loss": 0.189,
+      "step": 2070
+    },
+    {
+      "epoch": 5.5322687957418495,
+      "grad_norm": 6.797444820404053,
+      "learning_rate": 0.00016628180817610964,
+      "loss": 0.3289,
+      "step": 2075
+    },
+    {
+      "epoch": 5.545575515635396,
+      "grad_norm": 5.592912673950195,
+      "learning_rate": 0.00016612325167276394,
+      "loss": 0.3153,
+      "step": 2080
+    },
+    {
+      "epoch": 5.558882235528942,
+      "grad_norm": 2.010084629058838,
+      "learning_rate": 0.00016596439923132017,
+      "loss": 0.1268,
+      "step": 2085
+    },
+    {
+      "epoch": 5.572188955422488,
+      "grad_norm": 30.318086624145508,
+      "learning_rate": 0.0001658052515627308,
+      "loss": 0.3523,
+      "step": 2090
+    },
+    {
+      "epoch": 5.585495675316035,
+      "grad_norm": 27.016624450683594,
+      "learning_rate": 0.00016564580937926962,
+      "loss": 0.2108,
+      "step": 2095
+    },
+    {
+      "epoch": 5.598802395209581,
+      "grad_norm": 9.180904388427734,
+      "learning_rate": 0.00016548607339452853,
+      "loss": 0.0995,
+      "step": 2100
+    },
+    {
+      "epoch": 5.612109115103127,
+      "grad_norm": 0.2903202474117279,
+      "learning_rate": 0.0001653260443234143,
+      "loss": 0.0875,
+      "step": 2105
+    },
+    {
+      "epoch": 5.625415834996673,
+      "grad_norm": 3.65264630317688,
+      "learning_rate": 0.00016516572288214552,
+      "loss": 0.4259,
+      "step": 2110
+    },
+    {
+      "epoch": 5.63872255489022,
+      "grad_norm": 26.856353759765625,
+      "learning_rate": 0.00016500510978824926,
+      "loss": 0.176,
+      "step": 2115
+    },
+    {
+      "epoch": 5.652029274783766,
+      "grad_norm": 28.680461883544922,
+      "learning_rate": 0.00016484420576055785,
+      "loss": 0.2412,
+      "step": 2120
+    },
+    {
+      "epoch": 5.665335994677312,
+      "grad_norm": 2.5007455348968506,
+      "learning_rate": 0.00016468301151920575,
+      "loss": 0.1002,
+      "step": 2125
+    },
+    {
+      "epoch": 5.6786427145708585,
+      "grad_norm": 13.849712371826172,
+      "learning_rate": 0.0001645215277856263,
+      "loss": 0.1694,
+      "step": 2130
+    },
+    {
+      "epoch": 5.691949434464404,
+      "grad_norm": 0.9836284518241882,
+      "learning_rate": 0.0001643597552825485,
+      "loss": 0.4659,
+      "step": 2135
+    },
+    {
+      "epoch": 5.705256154357951,
+      "grad_norm": 6.066076755523682,
+      "learning_rate": 0.00016419769473399363,
+      "loss": 0.2144,
+      "step": 2140
+    },
+    {
+      "epoch": 5.718562874251497,
+      "grad_norm": 0.6988075375556946,
+      "learning_rate": 0.00016403534686527225,
+      "loss": 0.3341,
+      "step": 2145
+    },
+    {
+      "epoch": 5.731869594145043,
+      "grad_norm": 1.4430186748504639,
+      "learning_rate": 0.0001638727124029808,
+      "loss": 0.117,
+      "step": 2150
+    },
+    {
+      "epoch": 5.74517631403859,
+      "grad_norm": 0.7749199867248535,
+      "learning_rate": 0.00016370979207499845,
+      "loss": 0.0893,
+      "step": 2155
+    },
+    {
+      "epoch": 5.758483033932135,
+      "grad_norm": 2.697474479675293,
+      "learning_rate": 0.00016354658661048364,
+      "loss": 0.2948,
+      "step": 2160
+    },
+    {
+      "epoch": 5.771789753825682,
+      "grad_norm": 0.0894465297460556,
+      "learning_rate": 0.00016338309673987101,
+      "loss": 0.1375,
+      "step": 2165
+    },
+    {
+      "epoch": 5.7850964737192285,
+      "grad_norm": 7.317739963531494,
+      "learning_rate": 0.0001632193231948682,
+      "loss": 0.1734,
+      "step": 2170
+    },
+    {
+      "epoch": 5.798403193612774,
+      "grad_norm": 2.099191665649414,
+      "learning_rate": 0.00016305526670845226,
+      "loss": 0.1395,
+      "step": 2175
     },
     {
-      "epoch": 19.752351097178682,
-      "step": 1580,
-      "total_flos": 8695327329615872.0,
-      "train_loss": 0.2892922113949427,
-      "train_runtime": 830.7771,
-      "train_samples_per_second": 122.777,
-      "train_steps_per_second": 1.902
+      "epoch": 5.811709913506321,
+      "grad_norm": 3.64398193359375,
+      "learning_rate": 0.00016289092801486667,
+      "loss": 0.1896,
+      "step": 2180
+    },
+    {
+      "epoch": 5.825016633399867,
+      "grad_norm": 3.6657869815826416,
+      "learning_rate": 0.00016272630784961787,
+      "loss": 0.1412,
+      "step": 2185
+    },
+    {
+      "epoch": 5.838323353293413,
+      "grad_norm": 1.3434737920761108,
+      "learning_rate": 0.00016256140694947217,
+      "loss": 0.1477,
+      "step": 2190
+    },
+    {
+      "epoch": 5.85163007318696,
+      "grad_norm": 0.9189901351928711,
+      "learning_rate": 0.00016239622605245216,
+      "loss": 0.1273,
+      "step": 2195
+    },
+    {
+      "epoch": 5.864936793080505,
+      "grad_norm": 42.35541534423828,
+      "learning_rate": 0.00016223076589783368,
+      "loss": 0.1507,
+      "step": 2200
+    },
+    {
+      "epoch": 5.878243512974052,
+      "grad_norm": 5.649895668029785,
+      "learning_rate": 0.00016206502722614238,
+      "loss": 0.2484,
+      "step": 2205
+    },
+    {
+      "epoch": 5.891550232867598,
+      "grad_norm": 1.5216484069824219,
+      "learning_rate": 0.00016189901077915043,
+      "loss": 0.0698,
+      "step": 2210
+    },
+    {
+      "epoch": 5.904856952761144,
+      "grad_norm": 18.50678825378418,
+      "learning_rate": 0.0001617327172998732,
+      "loss": 0.1503,
+      "step": 2215
+    },
+    {
+      "epoch": 5.918163672654691,
+      "grad_norm": 25.463354110717773,
+      "learning_rate": 0.0001615661475325658,
+      "loss": 0.2796,
+      "step": 2220
+    },
+    {
+      "epoch": 5.931470392548237,
+      "grad_norm": 0.9343916773796082,
+      "learning_rate": 0.0001613993022227202,
+      "loss": 0.4415,
+      "step": 2225
+    },
+    {
+      "epoch": 5.944777112441783,
+      "grad_norm": 0.4629737436771393,
+      "learning_rate": 0.00016123218211706126,
+      "loss": 0.2218,
+      "step": 2230
+    },
+    {
+      "epoch": 5.95808383233533,
+      "grad_norm": 4.891594886779785,
+      "learning_rate": 0.00016106478796354382,
+      "loss": 0.1428,
+      "step": 2235
+    },
+    {
+      "epoch": 5.9713905522288755,
+      "grad_norm": 0.6886902451515198,
+      "learning_rate": 0.00016089712051134926,
+      "loss": 0.1396,
+      "step": 2240
+    },
+    {
+      "epoch": 5.984697272122422,
+      "grad_norm": 25.898942947387695,
+      "learning_rate": 0.0001607291805108821,
+      "loss": 0.1493,
+      "step": 2245
+    },
+    {
+      "epoch": 5.998003992015968,
+      "grad_norm": 9.379130363464355,
+      "learning_rate": 0.00016056096871376667,
+      "loss": 0.2857,
+      "step": 2250
+    },
+    {
+      "epoch": 5.998003992015968,
+      "eval_loss": 0.8669371604919434,
+      "eval_macro_f1": 67.23263039823122,
+      "eval_macro_precision": 68.19497057722046,
+      "eval_macro_recall": 67.175028478983,
+      "eval_micro_f1": 86.78892215568862,
+      "eval_micro_precision": 86.78892215568862,
+      "eval_micro_recall": 86.78892215568862,
+      "eval_runtime": 4.5626,
+      "eval_samples_per_second": 585.632,
+      "eval_steps_per_second": 36.602,
+      "step": 2250
+    },
+    {
+      "epoch": 6.013306719893547,
+      "grad_norm": 0.7099379301071167,
+      "learning_rate": 0.00016039248587284373,
+      "loss": 0.1042,
+      "step": 2255
+    },
+    {
+      "epoch": 6.026613439787092,
+      "grad_norm": 0.3128296732902527,
+      "learning_rate": 0.0001602237327421671,
+      "loss": 0.1591,
+      "step": 2260
+    },
+    {
+      "epoch": 6.039920159680639,
+      "grad_norm": 0.34373024106025696,
+      "learning_rate": 0.00016005471007700031,
+      "loss": 0.0785,
+      "step": 2265
+    },
+    {
+      "epoch": 6.053226879574185,
+      "grad_norm": 0.020252332091331482,
+      "learning_rate": 0.00015988541863381323,
+      "loss": 0.1448,
+      "step": 2270
+    },
+    {
+      "epoch": 6.066533599467731,
+      "grad_norm": 0.09661275893449783,
+      "learning_rate": 0.00015971585917027862,
+      "loss": 0.0803,
+      "step": 2275
+    },
+    {
+      "epoch": 6.079840319361278,
+      "grad_norm": 49.84988021850586,
+      "learning_rate": 0.0001595460324452688,
+      "loss": 0.1813,
+      "step": 2280
+    },
+    {
+      "epoch": 6.0931470392548235,
+      "grad_norm": 0.08736708760261536,
+      "learning_rate": 0.00015937593921885225,
+      "loss": 0.0707,
+      "step": 2285
+    },
+    {
+      "epoch": 6.10645375914837,
+      "grad_norm": 38.021400451660156,
+      "learning_rate": 0.00015920558025229014,
+      "loss": 0.2836,
+      "step": 2290
+    },
+    {
+      "epoch": 6.119760479041916,
+      "grad_norm": 0.023501181975007057,
+      "learning_rate": 0.000159034956308033,
+      "loss": 0.081,
+      "step": 2295
+    },
+    {
+      "epoch": 6.133067198935462,
+      "grad_norm": 6.965376853942871,
+      "learning_rate": 0.00015886406814971728,
+      "loss": 0.2449,
+      "step": 2300
+    },
+    {
+      "epoch": 6.146373918829009,
+      "grad_norm": 24.30607032775879,
+      "learning_rate": 0.000158692916542162,
+      "loss": 0.1398,
+      "step": 2305
+    },
+    {
+      "epoch": 6.159680638722555,
+      "grad_norm": 3.2971627712249756,
+      "learning_rate": 0.00015852150225136518,
+      "loss": 0.246,
+      "step": 2310
+    },
+    {
+      "epoch": 6.172987358616101,
+      "grad_norm": 0.4873849153518677,
+      "learning_rate": 0.00015834982604450045,
+      "loss": 0.1162,
+      "step": 2315
+    },
+    {
+      "epoch": 6.186294078509647,
+      "grad_norm": 21.559432983398438,
+      "learning_rate": 0.00015817788868991377,
+      "loss": 0.2077,
+      "step": 2320
+    },
+    {
+      "epoch": 6.199600798403194,
+      "grad_norm": 47.79915237426758,
+      "learning_rate": 0.00015800569095711982,
+      "loss": 0.2076,
+      "step": 2325
+    },
+    {
+      "epoch": 6.21290751829674,
+      "grad_norm": 11.995939254760742,
+      "learning_rate": 0.00015783323361679864,
+      "loss": 0.2112,
+      "step": 2330
+    },
+    {
+      "epoch": 6.226214238190286,
+      "grad_norm": 5.725085258483887,
+      "learning_rate": 0.00015766051744079218,
+      "loss": 0.3894,
+      "step": 2335
+    },
+    {
+      "epoch": 6.2395209580838324,
+      "grad_norm": 22.34183120727539,
+      "learning_rate": 0.00015748754320210072,
+      "loss": 0.1221,
+      "step": 2340
+    },
+    {
+      "epoch": 6.252827677977379,
+      "grad_norm": 8.4489107131958,
+      "learning_rate": 0.00015731431167487965,
+      "loss": 0.1507,
+      "step": 2345
+    },
+    {
+      "epoch": 6.266134397870925,
+      "grad_norm": 9.290181159973145,
+      "learning_rate": 0.00015714082363443575,
+      "loss": 0.2869,
+      "step": 2350
+    },
+    {
+      "epoch": 6.279441117764471,
+      "grad_norm": 3.636052131652832,
+      "learning_rate": 0.0001569670798572239,
+      "loss": 0.226,
+      "step": 2355
+    },
+    {
+      "epoch": 6.292747837658017,
+      "grad_norm": 29.546716690063477,
+      "learning_rate": 0.0001567930811208435,
+      "loss": 0.1516,
+      "step": 2360
+    },
+    {
+      "epoch": 6.306054557551564,
+      "grad_norm": 6.995582580566406,
+      "learning_rate": 0.00015661882820403516,
+      "loss": 0.0518,
+      "step": 2365
+    },
+    {
+      "epoch": 6.319361277445109,
+      "grad_norm": 4.880093574523926,
+      "learning_rate": 0.00015644432188667695,
+      "loss": 0.0536,
+      "step": 2370
+    },
+    {
+      "epoch": 6.332667997338656,
+      "grad_norm": 0.18312983214855194,
+      "learning_rate": 0.00015626956294978103,
+      "loss": 0.0382,
+      "step": 2375
+    },
+    {
+      "epoch": 6.3459747172322025,
+      "grad_norm": 0.03832171857357025,
+      "learning_rate": 0.00015609455217549032,
+      "loss": 0.156,
+      "step": 2380
+    },
+    {
+      "epoch": 6.359281437125748,
+      "grad_norm": 47.08546829223633,
+      "learning_rate": 0.0001559192903470747,
+      "loss": 0.2537,
+      "step": 2385
+    },
+    {
+      "epoch": 6.372588157019295,
+      "grad_norm": 26.35301399230957,
+      "learning_rate": 0.00015574377824892777,
+      "loss": 0.1868,
+      "step": 2390
+    },
+    {
+      "epoch": 6.385894876912841,
+      "grad_norm": 10.239131927490234,
+      "learning_rate": 0.00015556801666656312,
+      "loss": 0.4423,
+      "step": 2395
+    },
+    {
+      "epoch": 6.399201596806387,
+      "grad_norm": 0.012938698753714561,
+      "learning_rate": 0.00015539200638661104,
+      "loss": 0.0325,
+      "step": 2400
+    },
+    {
+      "epoch": 6.412508316699934,
+      "grad_norm": 0.37707385420799255,
+      "learning_rate": 0.0001552157481968148,
+      "loss": 0.12,
+      "step": 2405
+    },
+    {
+      "epoch": 6.425815036593479,
+      "grad_norm": 31.20591926574707,
+      "learning_rate": 0.00015503924288602715,
+      "loss": 0.0575,
+      "step": 2410
+    },
+    {
+      "epoch": 6.439121756487026,
+      "grad_norm": 16.725509643554688,
+      "learning_rate": 0.000154862491244207,
+      "loss": 0.3234,
+      "step": 2415
+    },
+    {
+      "epoch": 6.452428476380573,
+      "grad_norm": 2.5812032222747803,
+      "learning_rate": 0.0001546854940624156,
+      "loss": 0.1438,
+      "step": 2420
+    },
+    {
+      "epoch": 6.465735196274118,
+      "grad_norm": 7.374899864196777,
+      "learning_rate": 0.00015450825213281317,
+      "loss": 0.3837,
+      "step": 2425
+    },
+    {
+      "epoch": 6.479041916167665,
+      "grad_norm": 0.6281996369361877,
+      "learning_rate": 0.00015433076624865531,
+      "loss": 0.2103,
+      "step": 2430
+    },
+    {
+      "epoch": 6.492348636061211,
+      "grad_norm": 8.941553115844727,
+      "learning_rate": 0.00015415303720428945,
+      "loss": 0.5357,
+      "step": 2435
+    },
+    {
+      "epoch": 6.505655355954757,
+      "grad_norm": 0.4060121774673462,
+      "learning_rate": 0.0001539750657951513,
+      "loss": 0.2006,
+      "step": 2440
+    },
+    {
+      "epoch": 6.518962075848304,
+      "grad_norm": 30.6502628326416,
+      "learning_rate": 0.00015379685281776125,
+      "loss": 0.2649,
+      "step": 2445
+    },
+    {
+      "epoch": 6.5322687957418495,
+      "grad_norm": 0.9821872711181641,
+      "learning_rate": 0.00015361839906972096,
+      "loss": 0.1418,
+      "step": 2450
+    },
+    {
+      "epoch": 6.545575515635396,
+      "grad_norm": 2.888770341873169,
+      "learning_rate": 0.00015343970534970947,
+      "loss": 0.2907,
+      "step": 2455
+    },
+    {
+      "epoch": 6.558882235528942,
+      "grad_norm": 0.7967215180397034,
+      "learning_rate": 0.00015326077245747999,
+      "loss": 0.0516,
+      "step": 2460
+    },
+    {
+      "epoch": 6.572188955422488,
+      "grad_norm": 38.39204788208008,
+      "learning_rate": 0.00015308160119385615,
+      "loss": 0.154,
+      "step": 2465
+    },
+    {
+      "epoch": 6.585495675316035,
+      "grad_norm": 1.5320662260055542,
+      "learning_rate": 0.00015290219236072835,
+      "loss": 0.0774,
+      "step": 2470
+    },
+    {
+      "epoch": 6.598802395209581,
+      "grad_norm": 0.0738963857293129,
+      "learning_rate": 0.00015272254676105025,
+      "loss": 0.0356,
+      "step": 2475
+    },
+    {
+      "epoch": 6.612109115103127,
+      "grad_norm": 53.34037399291992,
+      "learning_rate": 0.00015254266519883525,
+      "loss": 0.2408,
+      "step": 2480
+    },
+    {
+      "epoch": 6.625415834996673,
+      "grad_norm": 1.6895263195037842,
+      "learning_rate": 0.00015236254847915274,
+      "loss": 0.1396,
+      "step": 2485
+    },
+    {
+      "epoch": 6.63872255489022,
+      "grad_norm": 2.0314857959747314,
+      "learning_rate": 0.0001521821974081246,
+      "loss": 0.0907,
+      "step": 2490
+    },
+    {
+      "epoch": 6.652029274783766,
+      "grad_norm": 6.590915679931641,
+      "learning_rate": 0.00015200161279292155,
+      "loss": 0.1152,
+      "step": 2495
+    },
+    {
+      "epoch": 6.665335994677312,
+      "grad_norm": 11.798026084899902,
+      "learning_rate": 0.00015182079544175955,
+      "loss": 0.0845,
+      "step": 2500
+    },
+    {
+      "epoch": 6.6786427145708585,
+      "grad_norm": 2.838351011276245,
+      "learning_rate": 0.0001516397461638962,
+      "loss": 0.3289,
+      "step": 2505
+    },
+    {
+      "epoch": 6.691949434464404,
+      "grad_norm": 1.8574706315994263,
+      "learning_rate": 0.0001514584657696271,
+      "loss": 0.1179,
+      "step": 2510
+    },
+    {
+      "epoch": 6.705256154357951,
+      "grad_norm": 2.1909406185150146,
+      "learning_rate": 0.00015127695507028213,
+      "loss": 0.2805,
+      "step": 2515
+    },
+    {
+      "epoch": 6.718562874251497,
+      "grad_norm": 0.1210150420665741,
+      "learning_rate": 0.00015109521487822206,
+      "loss": 0.0867,
+      "step": 2520
+    },
+    {
+      "epoch": 6.731869594145043,
+      "grad_norm": 14.33542537689209,
+      "learning_rate": 0.00015091324600683472,
+      "loss": 0.2439,
+      "step": 2525
+    },
+    {
+      "epoch": 6.74517631403859,
+      "grad_norm": 26.803274154663086,
+      "learning_rate": 0.0001507310492705313,
+      "loss": 0.1219,
+      "step": 2530
+    },
+    {
+      "epoch": 6.758483033932135,
+      "grad_norm": 0.3590647578239441,
+      "learning_rate": 0.000150548625484743,
+      "loss": 0.193,
+      "step": 2535
+    },
+    {
+      "epoch": 6.771789753825682,
+      "grad_norm": 0.17284749448299408,
+      "learning_rate": 0.00015036597546591699,
+      "loss": 0.0963,
+      "step": 2540
+    },
+    {
+      "epoch": 6.7850964737192285,
+      "grad_norm": 2.4201717376708984,
+      "learning_rate": 0.00015018310003151312,
+      "loss": 0.1209,
+      "step": 2545
+    },
+    {
+      "epoch": 6.798403193612774,
+      "grad_norm": 1.2542929649353027,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 0.2455,
+      "step": 2550
+    },
+    {
+      "epoch": 6.811709913506321,
+      "grad_norm": 0.8970535397529602,
+      "learning_rate": 0.0001498166761908515,
+      "loss": 0.0261,
+      "step": 2555
+    },
+    {
+      "epoch": 6.825016633399867,
+      "grad_norm": 3.8939568996429443,
+      "learning_rate": 0.000149633129424543,
+      "loss": 0.3337,
+      "step": 2560
+    },
+    {
+      "epoch": 6.838323353293413,
+      "grad_norm": 14.508380889892578,
+      "learning_rate": 0.0001494493605225477,
+      "loss": 0.1373,
+      "step": 2565
+    },
+    {
+      "epoch": 6.85163007318696,
+      "grad_norm": 0.10055989772081375,
+      "learning_rate": 0.00014926537030733302,
+      "loss": 0.1465,
+      "step": 2570
+    },
+    {
+      "epoch": 6.864936793080505,
+      "grad_norm": 0.039826489984989166,
+      "learning_rate": 0.00014908115960235682,
+      "loss": 0.1221,
+      "step": 2575
+    },
+    {
+      "epoch": 6.878243512974052,
+      "grad_norm": 0.798770010471344,
+      "learning_rate": 0.0001488967292320639,
+      "loss": 0.118,
+      "step": 2580
+    },
+    {
+      "epoch": 6.891550232867598,
+      "grad_norm": 0.03231286630034447,
+      "learning_rate": 0.00014871208002188203,
+      "loss": 0.1685,
+      "step": 2585
+    },
+    {
+      "epoch": 6.904856952761144,
+      "grad_norm": 42.45045471191406,
+      "learning_rate": 0.00014852721279821852,
+      "loss": 0.3227,
+      "step": 2590
+    },
+    {
+      "epoch": 6.918163672654691,
+      "grad_norm": 19.435428619384766,
+      "learning_rate": 0.00014834212838845637,
+      "loss": 0.0843,
+      "step": 2595
+    },
+    {
+      "epoch": 6.931470392548237,
+      "grad_norm": 26.672100067138672,
+      "learning_rate": 0.00014815682762095065,
+      "loss": 0.1584,
+      "step": 2600
+    },
+    {
+      "epoch": 6.944777112441783,
+      "grad_norm": 13.749527931213379,
+      "learning_rate": 0.00014797131132502465,
+      "loss": 0.1462,
+      "step": 2605
+    },
+    {
+      "epoch": 6.95808383233533,
+      "grad_norm": 0.3556053638458252,
+      "learning_rate": 0.00014778558033096633,
+      "loss": 0.5992,
+      "step": 2610
+    },
+    {
+      "epoch": 6.9713905522288755,
+      "grad_norm": 1.5534833669662476,
+      "learning_rate": 0.00014759963547002458,
+      "loss": 0.369,
+      "step": 2615
+    },
+    {
+      "epoch": 6.984697272122422,
+      "grad_norm": 0.774702787399292,
+      "learning_rate": 0.0001474134775744054,
+      "loss": 0.2468,
+      "step": 2620
+    },
+    {
+      "epoch": 6.998003992015968,
+      "grad_norm": 3.043125867843628,
+      "learning_rate": 0.0001472271074772683,
+      "loss": 0.1326,
+      "step": 2625
+    },
+    {
+      "epoch": 6.998003992015968,
+      "eval_loss": 0.7454941868782043,
+      "eval_macro_f1": 68.01755519864683,
+      "eval_macro_precision": 69.74014647112368,
+      "eval_macro_recall": 67.37265826177689,
+      "eval_micro_f1": 87.12574850299401,
+      "eval_micro_precision": 87.12574850299401,
+      "eval_micro_recall": 87.12574850299401,
+      "eval_runtime": 4.9043,
+      "eval_samples_per_second": 544.832,
+      "eval_steps_per_second": 34.052,
+      "step": 2625
+    },
+    {
+      "epoch": 7.013306719893547,
+      "grad_norm": 3.137073516845703,
+      "learning_rate": 0.00014704052601272242,
+      "loss": 0.0919,
+      "step": 2630
+    },
+    {
+      "epoch": 7.026613439787092,
+      "grad_norm": 0.6686985492706299,
+      "learning_rate": 0.00014685373401582296,
+      "loss": 0.0532,
+      "step": 2635
+    },
+    {
+      "epoch": 7.039920159680639,
+      "grad_norm": 0.02329118922352791,
+      "learning_rate": 0.00014666673232256738,
+      "loss": 0.0643,
+      "step": 2640
+    },
+    {
+      "epoch": 7.053226879574185,
+      "grad_norm": 1.519888997077942,
+      "learning_rate": 0.0001464795217698916,
+      "loss": 0.0942,
+      "step": 2645
+    },
+    {
+      "epoch": 7.066533599467731,
+      "grad_norm": 4.10860538482666,
+      "learning_rate": 0.00014629210319566627,
+      "loss": 0.0289,
+      "step": 2650
+    },
+    {
+      "epoch": 7.079840319361278,
+      "grad_norm": 3.4810853004455566,
+      "learning_rate": 0.00014610447743869314,
+      "loss": 0.3145,
+      "step": 2655
+    },
+    {
+      "epoch": 7.0931470392548235,
+      "grad_norm": 3.5535147190093994,
+      "learning_rate": 0.00014591664533870118,
+      "loss": 0.089,
+      "step": 2660
+    },
+    {
+      "epoch": 7.10645375914837,
+      "grad_norm": 0.7342696785926819,
+      "learning_rate": 0.00014572860773634286,
+      "loss": 0.0405,
+      "step": 2665
+    },
+    {
+      "epoch": 7.119760479041916,
+      "grad_norm": 0.0036852105986326933,
+      "learning_rate": 0.00014554036547319033,
+      "loss": 0.1064,
+      "step": 2670
+    },
+    {
+      "epoch": 7.133067198935462,
+      "grad_norm": 0.3384684920310974,
+      "learning_rate": 0.00014535191939173177,
+      "loss": 0.1589,
+      "step": 2675
+    },
+    {
+      "epoch": 7.146373918829009,
+      "grad_norm": 24.363168716430664,
+      "learning_rate": 0.0001451632703353676,
+      "loss": 0.1389,
+      "step": 2680
+    },
+    {
+      "epoch": 7.159680638722555,
+      "grad_norm": 0.233006089925766,
+      "learning_rate": 0.0001449744191484066,
+      "loss": 0.2027,
+      "step": 2685
+    },
+    {
+      "epoch": 7.172987358616101,
+      "grad_norm": 0.4666266143321991,
+      "learning_rate": 0.00014478536667606218,
+      "loss": 0.1332,
+      "step": 2690
+    },
+    {
+      "epoch": 7.186294078509647,
+      "grad_norm": 1.2150986194610596,
+      "learning_rate": 0.00014459611376444864,
+      "loss": 0.1118,
+      "step": 2695
+    },
+    {
+      "epoch": 7.199600798403194,
+      "grad_norm": 2.3224456310272217,
+      "learning_rate": 0.00014440666126057744,
+      "loss": 0.2385,
+      "step": 2700
+    },
+    {
+      "epoch": 7.21290751829674,
+      "grad_norm": 2.5107779502868652,
+      "learning_rate": 0.00014421701001235315,
+      "loss": 0.1927,
+      "step": 2705
+    },
+    {
+      "epoch": 7.226214238190286,
+      "grad_norm": 11.970067024230957,
+      "learning_rate": 0.00014402716086856998,
+      "loss": 0.1681,
+      "step": 2710
+    },
+    {
+      "epoch": 7.2395209580838324,
+      "grad_norm": 2.186711072921753,
+      "learning_rate": 0.00014383711467890774,
+      "loss": 0.1937,
+      "step": 2715
+    },
+    {
+      "epoch": 7.252827677977379,
+      "grad_norm": 0.2929200828075409,
+      "learning_rate": 0.00014364687229392824,
+      "loss": 0.0433,
+      "step": 2720
+    },
+    {
+      "epoch": 7.266134397870925,
+      "grad_norm": 1.6432567834854126,
+      "learning_rate": 0.00014345643456507124,
+      "loss": 0.288,
+      "step": 2725
+    },
+    {
+      "epoch": 7.279441117764471,
+      "grad_norm": 17.002174377441406,
+      "learning_rate": 0.00014326580234465085,
+      "loss": 0.1176,
+      "step": 2730
+    },
+    {
+      "epoch": 7.292747837658017,
+      "grad_norm": 1.520616888999939,
+      "learning_rate": 0.00014307497648585163,
+      "loss": 0.1465,
+      "step": 2735
+    },
+    {
+      "epoch": 7.306054557551564,
+      "grad_norm": 1.1925784349441528,
+      "learning_rate": 0.0001428839578427247,
+      "loss": 0.1364,
+      "step": 2740
+    },
+    {
+      "epoch": 7.319361277445109,
+      "grad_norm": 0.4294731616973877,
+      "learning_rate": 0.0001426927472701842,
+      "loss": 0.1293,
+      "step": 2745
+    },
+    {
+      "epoch": 7.332667997338656,
+      "grad_norm": 1.5070288181304932,
+      "learning_rate": 0.000142501345624003,
+      "loss": 0.074,
+      "step": 2750
+    },
+    {
+      "epoch": 7.3459747172322025,
+      "grad_norm": 0.8141427636146545,
+      "learning_rate": 0.00014230975376080935,
+      "loss": 0.1783,
+      "step": 2755
+    },
+    {
+      "epoch": 7.359281437125748,
+      "grad_norm": 20.88873291015625,
+      "learning_rate": 0.00014211797253808268,
+      "loss": 0.2202,
+      "step": 2760
+    },
+    {
+      "epoch": 7.372588157019295,
+      "grad_norm": 29.461524963378906,
+      "learning_rate": 0.00014192600281414994,
+      "loss": 0.3192,
+      "step": 2765
+    },
+    {
+      "epoch": 7.385894876912841,
+      "grad_norm": 0.10030537843704224,
+      "learning_rate": 0.0001417338454481818,
+      "loss": 0.2397,
+      "step": 2770
+    },
+    {
+      "epoch": 7.399201596806387,
+      "grad_norm": 0.23451067507266998,
+      "learning_rate": 0.00014154150130018866,
+      "loss": 0.0601,
+      "step": 2775
+    },
+    {
+      "epoch": 7.412508316699934,
+      "grad_norm": 3.4268810749053955,
+      "learning_rate": 0.00014134897123101688,
+      "loss": 0.0454,
+      "step": 2780
+    },
+    {
+      "epoch": 7.425815036593479,
+      "grad_norm": 32.5483283996582,
+      "learning_rate": 0.00014115625610234495,
+      "loss": 0.2532,
+      "step": 2785
+    },
+    {
+      "epoch": 7.439121756487026,
+      "grad_norm": 0.9835365414619446,
+      "learning_rate": 0.00014096335677667954,
+      "loss": 0.1134,
+      "step": 2790
+    },
+    {
+      "epoch": 7.452428476380573,
+      "grad_norm": 4.208250522613525,
+      "learning_rate": 0.00014077027411735183,
+      "loss": 0.0956,
+      "step": 2795
+    },
+    {
+      "epoch": 7.465735196274118,
+      "grad_norm": 3.528183937072754,
+      "learning_rate": 0.0001405770089885134,
+      "loss": 0.1693,
+      "step": 2800
+    },
+    {
+      "epoch": 7.479041916167665,
+      "grad_norm": 5.837916851043701,
+      "learning_rate": 0.00014038356225513248,
+      "loss": 0.1955,
+      "step": 2805
+    },
+    {
+      "epoch": 7.492348636061211,
+      "grad_norm": 11.988131523132324,
+      "learning_rate": 0.00014018993478299017,
+      "loss": 0.1545,
+      "step": 2810
+    },
+    {
+      "epoch": 7.505655355954757,
+      "grad_norm": 0.6364012360572815,
+      "learning_rate": 0.00013999612743867643,
+      "loss": 0.1219,
+      "step": 2815
+    },
+    {
+      "epoch": 7.518962075848304,
+      "grad_norm": 0.8442111611366272,
+      "learning_rate": 0.00013980214108958624,
+      "loss": 0.2948,
+      "step": 2820
+    },
+    {
+      "epoch": 7.5322687957418495,
+      "grad_norm": 4.432058334350586,
+      "learning_rate": 0.0001396079766039157,
+      "loss": 0.0897,
+      "step": 2825
+    },
+    {
+      "epoch": 7.545575515635396,
+      "grad_norm": 31.221689224243164,
+      "learning_rate": 0.00013941363485065822,
+      "loss": 0.2486,
+      "step": 2830
+    },
+    {
+      "epoch": 7.558882235528942,
+      "grad_norm": 0.8469045162200928,
+      "learning_rate": 0.00013921911669960055,
+      "loss": 0.0861,
+      "step": 2835
+    },
+    {
+      "epoch": 7.572188955422488,
+      "grad_norm": 1.0864439010620117,
+      "learning_rate": 0.00013902442302131894,
+      "loss": 0.1102,
+      "step": 2840
+    },
+    {
+      "epoch": 7.585495675316035,
+      "grad_norm": 42.718570709228516,
+      "learning_rate": 0.00013882955468717524,
+      "loss": 0.0893,
+      "step": 2845
+    },
+    {
+      "epoch": 7.598802395209581,
+      "grad_norm": 7.022069931030273,
+      "learning_rate": 0.00013863451256931287,
+      "loss": 0.1172,
+      "step": 2850
+    },
+    {
+      "epoch": 7.612109115103127,
+      "grad_norm": 0.48820632696151733,
+      "learning_rate": 0.0001384392975406532,
+      "loss": 0.0843,
+      "step": 2855
+    },
+    {
+      "epoch": 7.625415834996673,
+      "grad_norm": 0.0070191072300076485,
+      "learning_rate": 0.00013824391047489128,
+      "loss": 0.1253,
+      "step": 2860
+    },
+    {
+      "epoch": 7.63872255489022,
+      "grad_norm": 0.38560786843299866,
+      "learning_rate": 0.0001380483522464923,
+      "loss": 0.0797,
+      "step": 2865
+    },
+    {
+      "epoch": 7.652029274783766,
+      "grad_norm": 5.101590156555176,
+      "learning_rate": 0.0001378526237306874,
+      "loss": 0.2165,
+      "step": 2870
+    },
+    {
+      "epoch": 7.665335994677312,
+      "grad_norm": 4.549017429351807,
+      "learning_rate": 0.00013765672580346987,
+      "loss": 0.5324,
+      "step": 2875
+    },
+    {
+      "epoch": 7.6786427145708585,
+      "grad_norm": 0.4050997495651245,
+      "learning_rate": 0.00013746065934159123,
+      "loss": 0.1377,
+      "step": 2880
+    },
+    {
+      "epoch": 7.691949434464404,
+      "grad_norm": 0.03987701237201691,
+      "learning_rate": 0.0001372644252225572,
+      "loss": 0.0435,
+      "step": 2885
+    },
+    {
+      "epoch": 7.705256154357951,
+      "grad_norm": 13.718110084533691,
+      "learning_rate": 0.00013706802432462395,
+      "loss": 0.4619,
+      "step": 2890
+    },
+    {
+      "epoch": 7.718562874251497,
+      "grad_norm": 0.03140031918883324,
+      "learning_rate": 0.0001368714575267941,
+      "loss": 0.0112,
+      "step": 2895
+    },
+    {
+      "epoch": 7.731869594145043,
+      "grad_norm": 0.15146759152412415,
+      "learning_rate": 0.00013667472570881264,
+      "loss": 0.0316,
+      "step": 2900
+    },
+    {
+      "epoch": 7.74517631403859,
+      "grad_norm": 0.6249554753303528,
+      "learning_rate": 0.00013647782975116326,
+      "loss": 0.1377,
+      "step": 2905
+    },
+    {
+      "epoch": 7.758483033932135,
+      "grad_norm": 0.06544558703899384,
+      "learning_rate": 0.0001362807705350641,
+      "loss": 0.1871,
+      "step": 2910
+    },
+    {
+      "epoch": 7.771789753825682,
+      "grad_norm": 3.160745143890381,
+      "learning_rate": 0.0001360835489424642,
+      "loss": 0.1802,
+      "step": 2915
+    },
+    {
+      "epoch": 7.7850964737192285,
+      "grad_norm": 13.635746002197266,
+      "learning_rate": 0.00013588616585603907,
+      "loss": 0.2655,
+      "step": 2920
+    },
+    {
+      "epoch": 7.798403193612774,
+      "grad_norm": 19.401350021362305,
+      "learning_rate": 0.00013568862215918717,
+      "loss": 0.1116,
+      "step": 2925
+    },
+    {
+      "epoch": 7.811709913506321,
+      "grad_norm": 0.4326021671295166,
+      "learning_rate": 0.00013549091873602578,
+      "loss": 0.1016,
+      "step": 2930
+    },
+    {
+      "epoch": 7.825016633399867,
+      "grad_norm": 2.293384075164795,
+      "learning_rate": 0.00013529305647138687,
+      "loss": 0.0957,
+      "step": 2935
+    },
+    {
+      "epoch": 7.838323353293413,
+      "grad_norm": 0.6911963224411011,
+      "learning_rate": 0.00013509503625081358,
+      "loss": 0.1346,
+      "step": 2940
+    },
+    {
+      "epoch": 7.85163007318696,
+      "grad_norm": 3.5838940143585205,
+      "learning_rate": 0.00013489685896055572,
+      "loss": 0.1863,
+      "step": 2945
+    },
+    {
+      "epoch": 7.864936793080505,
+      "grad_norm": 0.15642301738262177,
+      "learning_rate": 0.00013469852548756624,
+      "loss": 0.2231,
+      "step": 2950
+    },
+    {
+      "epoch": 7.878243512974052,
+      "grad_norm": 12.195527076721191,
+      "learning_rate": 0.00013450003671949706,
+      "loss": 0.1953,
+      "step": 2955
+    },
+    {
+      "epoch": 7.891550232867598,
+      "grad_norm": 2.0001468658447266,
+      "learning_rate": 0.00013430139354469515,
+      "loss": 0.1195,
+      "step": 2960
+    },
+    {
+      "epoch": 7.904856952761144,
+      "grad_norm": 11.265728950500488,
+      "learning_rate": 0.00013410259685219845,
+      "loss": 0.2233,
+      "step": 2965
+    },
+    {
+      "epoch": 7.918163672654691,
+      "grad_norm": 0.5948832631111145,
+      "learning_rate": 0.00013390364753173206,
+      "loss": 0.0816,
+      "step": 2970
+    },
+    {
+      "epoch": 7.931470392548237,
+      "grad_norm": 0.09405430406332016,
+      "learning_rate": 0.00013370454647370418,
+      "loss": 0.0419,
+      "step": 2975
+    },
+    {
+      "epoch": 7.944777112441783,
+      "grad_norm": 0.07850921154022217,
+      "learning_rate": 0.00013350529456920206,
+      "loss": 0.1481,
+      "step": 2980
+    },
+    {
+      "epoch": 7.95808383233533,
+      "grad_norm": 0.05034489929676056,
+      "learning_rate": 0.00013330589270998808,
+      "loss": 0.0068,
+      "step": 2985
+    },
+    {
+      "epoch": 7.9713905522288755,
+      "grad_norm": 2.067530632019043,
+      "learning_rate": 0.0001331063417884958,
+      "loss": 0.2244,
+      "step": 2990
+    },
+    {
+      "epoch": 7.984697272122422,
+      "grad_norm": 0.05471380800008774,
+      "learning_rate": 0.0001329066426978259,
+      "loss": 0.0941,
+      "step": 2995
+    },
+    {
+      "epoch": 7.998003992015968,
+      "grad_norm": 5.7279133796691895,
+      "learning_rate": 0.00013270679633174218,
+      "loss": 0.4085,
+      "step": 3000
+    },
+    {
+      "epoch": 7.998003992015968,
+      "eval_loss": 0.9578362107276917,
+      "eval_macro_f1": 67.62551780945887,
+      "eval_macro_precision": 73.99170552196239,
+      "eval_macro_recall": 64.56609836337587,
+      "eval_micro_f1": 88.02395209580838,
+      "eval_micro_precision": 88.02395209580838,
+      "eval_micro_recall": 88.02395209580838,
+      "eval_runtime": 4.5288,
+      "eval_samples_per_second": 590.002,
+      "eval_steps_per_second": 36.875,
+      "step": 3000
+    },
+    {
+      "epoch": 8.013306719893546,
+      "grad_norm": 23.61517333984375,
+      "learning_rate": 0.00013250680358466754,
+      "loss": 0.2144,
+      "step": 3005
+    },
+    {
+      "epoch": 8.026613439787093,
+      "grad_norm": 1.6236063241958618,
+      "learning_rate": 0.0001323066653516801,
+      "loss": 0.145,
+      "step": 3010
+    },
+    {
+      "epoch": 8.039920159680639,
+      "grad_norm": 0.12873251736164093,
+      "learning_rate": 0.00013210638252850908,
+      "loss": 0.0148,
+      "step": 3015
+    },
+    {
+      "epoch": 8.053226879574185,
+      "grad_norm": 1.4208223819732666,
+      "learning_rate": 0.0001319059560115308,
+      "loss": 0.1435,
+      "step": 3020
+    },
+    {
+      "epoch": 8.066533599467732,
+      "grad_norm": 1.1595903635025024,
+      "learning_rate": 0.00013170538669776468,
+      "loss": 0.0784,
+      "step": 3025
+    },
+    {
+      "epoch": 8.079840319361278,
+      "grad_norm": 2.5915064811706543,
+      "learning_rate": 0.0001315046754848693,
+      "loss": 0.092,
+      "step": 3030
+    },
+    {
+      "epoch": 8.093147039254823,
+      "grad_norm": 1.323926329612732,
+      "learning_rate": 0.00013130382327113823,
+      "loss": 0.0972,
+      "step": 3035
+    },
+    {
+      "epoch": 8.10645375914837,
+      "grad_norm": 1.1341009140014648,
+      "learning_rate": 0.00013110283095549614,
+      "loss": 0.0939,
+      "step": 3040
+    },
+    {
+      "epoch": 8.119760479041917,
+      "grad_norm": 0.8324944972991943,
+      "learning_rate": 0.00013090169943749476,
+      "loss": 0.1248,
+      "step": 3045
+    },
+    {
+      "epoch": 8.133067198935462,
+      "grad_norm": 0.12022148072719574,
+      "learning_rate": 0.00013070042961730877,
+      "loss": 0.0167,
+      "step": 3050
+    },
+    {
+      "epoch": 8.146373918829008,
+      "grad_norm": 0.08643687516450882,
+      "learning_rate": 0.00013049902239573187,
+      "loss": 0.0794,
+      "step": 3055
+    },
+    {
+      "epoch": 8.159680638722556,
+      "grad_norm": 0.017748689278960228,
+      "learning_rate": 0.00013029747867417276,
+      "loss": 0.0169,
+      "step": 3060
+    },
+    {
+      "epoch": 8.172987358616101,
+      "grad_norm": 0.03752296790480614,
+      "learning_rate": 0.00013009579935465085,
+      "loss": 0.0268,
+      "step": 3065
+    },
+    {
+      "epoch": 8.186294078509647,
+      "grad_norm": 0.5308043956756592,
+      "learning_rate": 0.0001298939853397927,
+      "loss": 0.1967,
+      "step": 3070
+    },
+    {
+      "epoch": 8.199600798403194,
+      "grad_norm": 2.3242459297180176,
+      "learning_rate": 0.0001296920375328275,
+      "loss": 0.0467,
+      "step": 3075
+    },
+    {
+      "epoch": 8.21290751829674,
+      "grad_norm": 0.10183367133140564,
+      "learning_rate": 0.0001294899568375833,
+      "loss": 0.0591,
+      "step": 3080
+    },
+    {
+      "epoch": 8.226214238190286,
+      "grad_norm": 0.013796583749353886,
+      "learning_rate": 0.00012928774415848295,
+      "loss": 0.1172,
+      "step": 3085
+    },
+    {
+      "epoch": 8.239520958083832,
+      "grad_norm": 0.060439061373472214,
+      "learning_rate": 0.0001290854004005399,
+      "loss": 0.1231,
+      "step": 3090
+    },
+    {
+      "epoch": 8.252827677977379,
+      "grad_norm": 0.04022917151451111,
+      "learning_rate": 0.0001288829264693544,
+      "loss": 0.0047,
+      "step": 3095
+    },
+    {
+      "epoch": 8.266134397870925,
+      "grad_norm": 1.164883017539978,
+      "learning_rate": 0.00012868032327110904,
+      "loss": 0.0269,
+      "step": 3100
+    },
+    {
+      "epoch": 8.27944111776447,
+      "grad_norm": 13.31337833404541,
+      "learning_rate": 0.00012847759171256523,
+      "loss": 0.1548,
+      "step": 3105
+    },
+    {
+      "epoch": 8.292747837658018,
+      "grad_norm": 0.5353259444236755,
+      "learning_rate": 0.00012827473270105873,
+      "loss": 0.0542,
+      "step": 3110
+    },
+    {
+      "epoch": 8.306054557551564,
+      "grad_norm": 0.004808424971997738,
+      "learning_rate": 0.00012807174714449571,
+      "loss": 0.0818,
+      "step": 3115
+    },
+    {
+      "epoch": 8.31936127744511,
+      "grad_norm": 0.9103612899780273,
+      "learning_rate": 0.0001278686359513488,
+      "loss": 0.0221,
+      "step": 3120
+    },
+    {
+      "epoch": 8.332667997338657,
+      "grad_norm": 48.99919128417969,
+      "learning_rate": 0.0001276654000306527,
+      "loss": 0.2673,
+      "step": 3125
+    },
+    {
+      "epoch": 8.345974717232203,
+      "grad_norm": 1.4264007806777954,
+      "learning_rate": 0.00012746204029200067,
+      "loss": 0.1625,
+      "step": 3130
+    },
+    {
+      "epoch": 8.359281437125748,
+      "grad_norm": 15.919877052307129,
+      "learning_rate": 0.0001272585576455398,
+      "loss": 0.1033,
+      "step": 3135
+    },
+    {
+      "epoch": 8.372588157019294,
+      "grad_norm": 49.93979263305664,
+      "learning_rate": 0.00012705495300196747,
+      "loss": 0.2639,
+      "step": 3140
+    },
+    {
+      "epoch": 8.385894876912841,
+      "grad_norm": 1.7972381114959717,
+      "learning_rate": 0.00012685122727252695,
+      "loss": 0.0508,
+      "step": 3145
+    },
+    {
+      "epoch": 8.399201596806387,
+      "grad_norm": 0.6764810085296631,
+      "learning_rate": 0.00012664738136900348,
+      "loss": 0.1182,
+      "step": 3150
+    },
+    {
+      "epoch": 8.412508316699933,
+      "grad_norm": 2.4136905670166016,
+      "learning_rate": 0.00012644341620372023,
+      "loss": 0.1769,
+      "step": 3155
+    },
+    {
+      "epoch": 8.42581503659348,
+      "grad_norm": 0.28830185532569885,
+      "learning_rate": 0.00012623933268953396,
+      "loss": 0.0763,
+      "step": 3160
+    },
+    {
+      "epoch": 8.439121756487026,
+      "grad_norm": 1.7211086750030518,
+      "learning_rate": 0.0001260351317398312,
+      "loss": 0.0386,
+      "step": 3165
+    },
+    {
+      "epoch": 8.452428476380572,
+      "grad_norm": 0.05553643777966499,
+      "learning_rate": 0.00012583081426852411,
+      "loss": 0.0522,
+      "step": 3170
+    },
+    {
+      "epoch": 8.46573519627412,
+      "grad_norm": 1.4411470890045166,
+      "learning_rate": 0.00012562638119004626,
+      "loss": 0.0946,
+      "step": 3175
+    },
+    {
+      "epoch": 8.479041916167665,
+      "grad_norm": 0.201587975025177,
+      "learning_rate": 0.00012542183341934872,
+      "loss": 0.0996,
+      "step": 3180
+    },
+    {
+      "epoch": 8.49234863606121,
+      "grad_norm": 1.4264519214630127,
+      "learning_rate": 0.00012521717187189573,
+      "loss": 0.2037,
+      "step": 3185
+    },
+    {
+      "epoch": 8.505655355954758,
+      "grad_norm": 2.3621273040771484,
+      "learning_rate": 0.00012501239746366092,
+      "loss": 0.0675,
+      "step": 3190
+    },
+    {
+      "epoch": 8.518962075848304,
+      "grad_norm": 1.260345458984375,
+      "learning_rate": 0.0001248075111111229,
+      "loss": 0.1723,
+      "step": 3195
+    },
+    {
+      "epoch": 8.53226879574185,
+      "grad_norm": 0.44682565331459045,
+      "learning_rate": 0.00012460251373126136,
+      "loss": 0.336,
+      "step": 3200
+    },
+    {
+      "epoch": 8.545575515635395,
+      "grad_norm": 0.35291796922683716,
+      "learning_rate": 0.00012439740624155284,
+      "loss": 0.0561,
+      "step": 3205
+    },
+    {
+      "epoch": 8.558882235528943,
+      "grad_norm": 0.7354090809822083,
+      "learning_rate": 0.00012419218955996676,
+      "loss": 0.2095,
+      "step": 3210
+    },
+    {
+      "epoch": 8.572188955422488,
+      "grad_norm": 2.9078779220581055,
+      "learning_rate": 0.00012398686460496122,
+      "loss": 0.0378,
+      "step": 3215
+    },
+    {
+      "epoch": 8.585495675316034,
+      "grad_norm": 0.8109474778175354,
+      "learning_rate": 0.00012378143229547882,
+      "loss": 0.236,
+      "step": 3220
+    },
+    {
+      "epoch": 8.598802395209582,
+      "grad_norm": 2.6716489791870117,
+      "learning_rate": 0.00012357589355094275,
+      "loss": 0.0557,
+      "step": 3225
+    },
+    {
+      "epoch": 8.612109115103127,
+      "grad_norm": 36.927249908447266,
+      "learning_rate": 0.00012337024929125242,
+      "loss": 0.1876,
+      "step": 3230
+    },
+    {
+      "epoch": 8.625415834996673,
+      "grad_norm": 1.459558129310608,
+      "learning_rate": 0.0001231645004367796,
+      "loss": 0.1989,
+      "step": 3235
+    },
+    {
+      "epoch": 8.638722554890219,
+      "grad_norm": 1.2060911655426025,
+      "learning_rate": 0.0001229586479083641,
+      "loss": 0.0587,
+      "step": 3240
+    },
+    {
+      "epoch": 8.652029274783766,
+      "grad_norm": 0.03878104314208031,
+      "learning_rate": 0.00012275269262730982,
+      "loss": 0.026,
+      "step": 3245
+    },
+    {
+      "epoch": 8.665335994677312,
+      "grad_norm": 0.032785095274448395,
+      "learning_rate": 0.00012254663551538046,
+      "loss": 0.1203,
+      "step": 3250
+    },
+    {
+      "epoch": 8.678642714570858,
+      "grad_norm": 0.01734250970184803,
+      "learning_rate": 0.00012234047749479544,
+      "loss": 0.0483,
+      "step": 3255
+    },
+    {
+      "epoch": 8.691949434464405,
+      "grad_norm": 0.004084521438926458,
+      "learning_rate": 0.0001221342194882259,
+      "loss": 0.1555,
+      "step": 3260
+    },
+    {
+      "epoch": 8.70525615435795,
+      "grad_norm": 0.4270322620868683,
+      "learning_rate": 0.00012192786241879033,
+      "loss": 0.0689,
+      "step": 3265
+    },
+    {
+      "epoch": 8.718562874251496,
+      "grad_norm": 73.75556945800781,
+      "learning_rate": 0.00012172140721005079,
+      "loss": 0.0641,
+      "step": 3270
+    },
+    {
+      "epoch": 8.731869594145044,
+      "grad_norm": 15.693498611450195,
+      "learning_rate": 0.00012151485478600839,
+      "loss": 0.2614,
+      "step": 3275
+    },
+    {
+      "epoch": 8.74517631403859,
+      "grad_norm": 0.04169188439846039,
+      "learning_rate": 0.00012130820607109936,
+      "loss": 0.0439,
+      "step": 3280
+    },
+    {
+      "epoch": 8.758483033932135,
+      "grad_norm": 0.4641147553920746,
+      "learning_rate": 0.000121101461990191,
+      "loss": 0.0238,
+      "step": 3285
+    },
+    {
+      "epoch": 8.771789753825683,
+      "grad_norm": 3.9536852836608887,
+      "learning_rate": 0.00012089462346857725,
+      "loss": 0.1257,
+      "step": 3290
+    },
+    {
+      "epoch": 8.785096473719229,
+      "grad_norm": 2.0397369861602783,
+      "learning_rate": 0.00012068769143197487,
+      "loss": 0.1766,
+      "step": 3295
+    },
+    {
+      "epoch": 8.798403193612774,
+      "grad_norm": 30.004472732543945,
+      "learning_rate": 0.00012048066680651908,
+      "loss": 0.0752,
+      "step": 3300
+    },
+    {
+      "epoch": 8.81170991350632,
+      "grad_norm": 1.0506658554077148,
+      "learning_rate": 0.0001202735505187595,
+      "loss": 0.2345,
+      "step": 3305
+    },
+    {
+      "epoch": 8.825016633399867,
+      "grad_norm": 1.2177345752716064,
+      "learning_rate": 0.00012006634349565602,
+      "loss": 0.0594,
+      "step": 3310
+    },
+    {
+      "epoch": 8.838323353293413,
+      "grad_norm": 0.415720134973526,
+      "learning_rate": 0.00011985904666457455,
+      "loss": 0.0283,
+      "step": 3315
+    },
+    {
+      "epoch": 8.851630073186959,
+      "grad_norm": 0.6650182604789734,
+      "learning_rate": 0.00011965166095328301,
+      "loss": 0.0981,
+      "step": 3320
+    },
+    {
+      "epoch": 8.864936793080506,
+      "grad_norm": 129.6624298095703,
+      "learning_rate": 0.00011944418728994709,
+      "loss": 0.1986,
+      "step": 3325
+    },
+    {
+      "epoch": 8.878243512974052,
+      "grad_norm": 18.93353843688965,
+      "learning_rate": 0.00011923662660312611,
+      "loss": 0.188,
+      "step": 3330
+    },
+    {
+      "epoch": 8.891550232867598,
+      "grad_norm": 0.8205522298812866,
+      "learning_rate": 0.00011902897982176882,
+      "loss": 0.0841,
+      "step": 3335
+    },
+    {
+      "epoch": 8.904856952761145,
+      "grad_norm": 32.96717071533203,
+      "learning_rate": 0.00011882124787520934,
+      "loss": 0.0548,
+      "step": 3340
+    },
+    {
+      "epoch": 8.918163672654691,
+      "grad_norm": 18.162229537963867,
+      "learning_rate": 0.00011861343169316301,
+      "loss": 0.0967,
+      "step": 3345
+    },
+    {
+      "epoch": 8.931470392548237,
+      "grad_norm": 0.480948269367218,
+      "learning_rate": 0.00011840553220572204,
+      "loss": 0.0964,
+      "step": 3350
+    },
+    {
+      "epoch": 8.944777112441782,
+      "grad_norm": 0.0540715716779232,
+      "learning_rate": 0.0001181975503433516,
+      "loss": 0.0041,
+      "step": 3355
+    },
+    {
+      "epoch": 8.95808383233533,
+      "grad_norm": 0.6189212799072266,
+      "learning_rate": 0.00011798948703688539,
+      "loss": 0.0134,
+      "step": 3360
+    },
+    {
+      "epoch": 8.971390552228875,
+      "grad_norm": 0.002311081625521183,
+      "learning_rate": 0.00011778134321752182,
+      "loss": 0.1133,
+      "step": 3365
+    },
+    {
+      "epoch": 8.984697272122421,
+      "grad_norm": 19.883926391601562,
+      "learning_rate": 0.00011757311981681942,
+      "loss": 0.072,
+      "step": 3370
+    },
+    {
+      "epoch": 8.998003992015969,
+      "grad_norm": 2.3699491024017334,
+      "learning_rate": 0.00011736481776669306,
+      "loss": 0.0273,
+      "step": 3375
+    },
+    {
+      "epoch": 8.998003992015969,
+      "eval_loss": 1.5414294004440308,
+      "eval_macro_f1": 64.80803991978306,
+      "eval_macro_precision": 70.6655091690804,
+      "eval_macro_recall": 61.210875301941044,
+      "eval_micro_f1": 87.12574850299401,
+      "eval_micro_precision": 87.12574850299401,
+      "eval_micro_recall": 87.12574850299401,
+      "eval_runtime": 5.1705,
+      "eval_samples_per_second": 516.777,
+      "eval_steps_per_second": 32.299,
+      "step": 3375
+    },
+    {
+      "epoch": 9.013306719893546,
+      "grad_norm": 0.21652141213417053,
+      "learning_rate": 0.0001171564379994095,
+      "loss": 0.0913,
+      "step": 3380
+    },
+    {
+      "epoch": 9.026613439787093,
+      "grad_norm": 0.7938780188560486,
+      "learning_rate": 0.0001169479814475834,
+      "loss": 0.0546,
+      "step": 3385
+    },
+    {
+      "epoch": 9.039920159680639,
+      "grad_norm": 0.004158825613558292,
+      "learning_rate": 0.00011673944904417308,
+      "loss": 0.102,
+      "step": 3390
+    },
+    {
+      "epoch": 9.053226879574185,
+      "grad_norm": 16.037822723388672,
+      "learning_rate": 0.00011653084172247624,
+      "loss": 0.3269,
+      "step": 3395
+    },
+    {
+      "epoch": 9.066533599467732,
+      "grad_norm": 12.818014144897461,
+      "learning_rate": 0.00011632216041612594,
+      "loss": 0.268,
+      "step": 3400
+    },
+    {
+      "epoch": 9.079840319361278,
+      "grad_norm": 0.016593078151345253,
+      "learning_rate": 0.00011611340605908642,
+      "loss": 0.01,
+      "step": 3405
+    },
+    {
+      "epoch": 9.093147039254823,
+      "grad_norm": 0.01048375479876995,
+      "learning_rate": 0.00011590457958564878,
+      "loss": 0.0739,
+      "step": 3410
+    },
+    {
+      "epoch": 9.10645375914837,
+      "grad_norm": 24.694929122924805,
+      "learning_rate": 0.0001156956819304269,
+      "loss": 0.0642,
+      "step": 3415
+    },
+    {
+      "epoch": 9.119760479041917,
+      "grad_norm": 38.852474212646484,
+      "learning_rate": 0.00011548671402835325,
+      "loss": 0.0525,
+      "step": 3420
+    },
+    {
+      "epoch": 9.133067198935462,
+      "grad_norm": 0.016515236347913742,
+      "learning_rate": 0.00011527767681467471,
+      "loss": 0.0399,
+      "step": 3425
+    },
+    {
+      "epoch": 9.146373918829008,
+      "grad_norm": 0.6999401450157166,
+      "learning_rate": 0.00011506857122494831,
+      "loss": 0.095,
+      "step": 3430
+    },
+    {
+      "epoch": 9.159680638722556,
+      "grad_norm": 0.023875270038843155,
+      "learning_rate": 0.00011485939819503717,
+      "loss": 0.0629,
+      "step": 3435
+    },
+    {
+      "epoch": 9.172987358616101,
+      "grad_norm": 0.0019072931027039886,
+      "learning_rate": 0.00011465015866110622,
+      "loss": 0.0449,
+      "step": 3440
+    },
+    {
+      "epoch": 9.186294078509647,
+      "grad_norm": 0.013696919195353985,
+      "learning_rate": 0.000114440853559618,
+      "loss": 0.0745,
+      "step": 3445
+    },
+    {
+      "epoch": 9.199600798403194,
+      "grad_norm": 0.05094515159726143,
+      "learning_rate": 0.00011423148382732853,
+      "loss": 0.0808,
+      "step": 3450
+    },
+    {
+      "epoch": 9.21290751829674,
+      "grad_norm": 0.33801838755607605,
+      "learning_rate": 0.00011402205040128307,
+      "loss": 0.0772,
+      "step": 3455
+    },
+    {
+      "epoch": 9.226214238190286,
+      "grad_norm": 6.502445220947266,
+      "learning_rate": 0.00011381255421881198,
+      "loss": 0.2097,
+      "step": 3460
+    },
+    {
+      "epoch": 9.239520958083832,
+      "grad_norm": 1.0928117036819458,
+      "learning_rate": 0.00011360299621752644,
+      "loss": 0.0446,
+      "step": 3465
+    },
+    {
+      "epoch": 9.252827677977379,
+      "grad_norm": 0.864954948425293,
+      "learning_rate": 0.00011339337733531434,
+      "loss": 0.088,
+      "step": 3470
+    },
+    {
+      "epoch": 9.266134397870925,
+      "grad_norm": 10.347973823547363,
+      "learning_rate": 0.00011318369851033603,
+      "loss": 0.0197,
+      "step": 3475
+    },
+    {
+      "epoch": 9.27944111776447,
+      "grad_norm": 1.25753653049469,
+      "learning_rate": 0.00011297396068102017,
+      "loss": 0.1665,
+      "step": 3480
+    },
+    {
+      "epoch": 9.292747837658018,
+      "grad_norm": 0.006565026007592678,
+      "learning_rate": 0.00011276416478605949,
+      "loss": 0.0072,
+      "step": 3485
+    },
+    {
+      "epoch": 9.306054557551564,
+      "grad_norm": 0.14825934171676636,
+      "learning_rate": 0.0001125543117644065,
+      "loss": 0.0801,
+      "step": 3490
+    },
+    {
+      "epoch": 9.31936127744511,
+      "grad_norm": 0.837486743927002,
+      "learning_rate": 0.00011234440255526948,
+      "loss": 0.1842,
+      "step": 3495
+    },
+    {
+      "epoch": 9.332667997338657,
+      "grad_norm": 4.4184064865112305,
+      "learning_rate": 0.0001121344380981082,
+      "loss": 0.1044,
+      "step": 3500
+    },
+    {
+      "epoch": 9.345974717232203,
+      "grad_norm": 0.4564765989780426,
+      "learning_rate": 0.00011192441933262962,
+      "loss": 0.0216,
+      "step": 3505
+    },
+    {
+      "epoch": 9.359281437125748,
+      "grad_norm": 4.210375785827637,
+      "learning_rate": 0.00011171434719878384,
+      "loss": 0.3018,
+      "step": 3510
+    },
+    {
+      "epoch": 9.372588157019294,
+      "grad_norm": 0.4331660866737366,
+      "learning_rate": 0.00011150422263675968,
+      "loss": 0.1116,
+      "step": 3515
+    },
+    {
+      "epoch": 9.385894876912841,
+      "grad_norm": 1.863437533378601,
+      "learning_rate": 0.00011129404658698081,
+      "loss": 0.1566,
+      "step": 3520
+    },
+    {
+      "epoch": 9.399201596806387,
+      "grad_norm": 1.0628738403320312,
+      "learning_rate": 0.00011108381999010111,
+      "loss": 0.0746,
+      "step": 3525
+    },
+    {
+      "epoch": 9.412508316699933,
+      "grad_norm": 1.0913702249526978,
+      "learning_rate": 0.00011087354378700086,
+      "loss": 0.0515,
+      "step": 3530
+    },
+    {
+      "epoch": 9.42581503659348,
+      "grad_norm": 0.40956100821495056,
+      "learning_rate": 0.00011066321891878227,
+      "loss": 0.0517,
+      "step": 3535
+    },
+    {
+      "epoch": 9.439121756487026,
+      "grad_norm": 31.597082138061523,
+      "learning_rate": 0.00011045284632676536,
+      "loss": 0.1389,
+      "step": 3540
+    },
+    {
+      "epoch": 9.452428476380572,
+      "grad_norm": 8.876389503479004,
+      "learning_rate": 0.00011024242695248379,
+      "loss": 0.1226,
+      "step": 3545
+    },
+    {
+      "epoch": 9.46573519627412,
+      "grad_norm": 0.767656683921814,
+      "learning_rate": 0.0001100319617376805,
+      "loss": 0.0444,
+      "step": 3550
+    },
+    {
+      "epoch": 9.479041916167665,
+      "grad_norm": 0.5712203979492188,
+      "learning_rate": 0.00010982145162430373,
+      "loss": 0.2447,
+      "step": 3555
+    },
+    {
+      "epoch": 9.49234863606121,
+      "grad_norm": 0.08574845641851425,
+      "learning_rate": 0.00010961089755450254,
+      "loss": 0.0436,
+      "step": 3560
+    },
+    {
+      "epoch": 9.505655355954758,
+      "grad_norm": 0.19157272577285767,
+      "learning_rate": 0.00010940030047062275,
+      "loss": 0.1011,
+      "step": 3565
+    },
+    {
+      "epoch": 9.518962075848304,
+      "grad_norm": 1.378658413887024,
+      "learning_rate": 0.00010918966131520277,
+      "loss": 0.0695,
+      "step": 3570
+    },
+    {
+      "epoch": 9.53226879574185,
+      "grad_norm": 0.7202726602554321,
+      "learning_rate": 0.00010897898103096917,
+      "loss": 0.1102,
+      "step": 3575
+    },
+    {
+      "epoch": 9.545575515635395,
+      "grad_norm": 0.020597225055098534,
+      "learning_rate": 0.00010876826056083273,
+      "loss": 0.0282,
+      "step": 3580
+    },
+    {
+      "epoch": 9.558882235528943,
+      "grad_norm": 0.7092624306678772,
+      "learning_rate": 0.00010855750084788398,
+      "loss": 0.1663,
+      "step": 3585
+    },
+    {
+      "epoch": 9.572188955422488,
+      "grad_norm": 0.05511481687426567,
+      "learning_rate": 0.00010834670283538914,
+      "loss": 0.1794,
+      "step": 3590
+    },
+    {
+      "epoch": 9.585495675316034,
+      "grad_norm": 3.0630874633789062,
+      "learning_rate": 0.00010813586746678583,
+      "loss": 0.1171,
+      "step": 3595
+    },
+    {
+      "epoch": 9.598802395209582,
+      "grad_norm": 0.7750560641288757,
+      "learning_rate": 0.00010792499568567884,
+      "loss": 0.0422,
+      "step": 3600
+    },
+    {
+      "epoch": 9.612109115103127,
+      "grad_norm": 0.020036980509757996,
+      "learning_rate": 0.00010771408843583598,
+      "loss": 0.0478,
+      "step": 3605
+    },
+    {
+      "epoch": 9.625415834996673,
+      "grad_norm": 0.4453740417957306,
+      "learning_rate": 0.0001075031466611837,
+      "loss": 0.0674,
+      "step": 3610
+    },
+    {
+      "epoch": 9.638722554890219,
+      "grad_norm": 0.08731366693973541,
+      "learning_rate": 0.0001072921713058031,
+      "loss": 0.0545,
+      "step": 3615
+    },
+    {
+      "epoch": 9.652029274783766,
+      "grad_norm": 0.90806645154953,
+      "learning_rate": 0.00010708116331392541,
+      "loss": 0.2005,
+      "step": 3620
+    },
+    {
+      "epoch": 9.665335994677312,
+      "grad_norm": 12.486852645874023,
+      "learning_rate": 0.0001068701236299281,
+      "loss": 0.0447,
+      "step": 3625
+    },
+    {
+      "epoch": 9.678642714570858,
+      "grad_norm": 0.44622567296028137,
+      "learning_rate": 0.00010665905319833041,
+      "loss": 0.0622,
+      "step": 3630
+    },
+    {
+      "epoch": 9.691949434464405,
+      "grad_norm": 0.10433344542980194,
+      "learning_rate": 0.00010644795296378909,
+      "loss": 0.0709,
+      "step": 3635
+    },
+    {
+      "epoch": 9.70525615435795,
+      "grad_norm": 0.2997910678386688,
+      "learning_rate": 0.00010623682387109447,
+      "loss": 0.2982,
+      "step": 3640
+    },
+    {
+      "epoch": 9.718562874251496,
+      "grad_norm": 1.653448462486267,
+      "learning_rate": 0.00010602566686516586,
+      "loss": 0.2432,
+      "step": 3645
+    },
+    {
+      "epoch": 9.731869594145044,
+      "grad_norm": 97.50984954833984,
+      "learning_rate": 0.00010581448289104758,
+      "loss": 0.3226,
+      "step": 3650
+    },
+    {
+      "epoch": 9.74517631403859,
+      "grad_norm": 0.7525721192359924,
+      "learning_rate": 0.00010560327289390468,
+      "loss": 0.0432,
+      "step": 3655
+    },
+    {
+      "epoch": 9.758483033932135,
+      "grad_norm": 1.433956503868103,
+      "learning_rate": 0.00010539203781901861,
+      "loss": 0.0749,
+      "step": 3660
+    },
+    {
+      "epoch": 9.771789753825683,
+      "grad_norm": 10.341765403747559,
+      "learning_rate": 0.00010518077861178309,
+      "loss": 0.0207,
+      "step": 3665
+    },
+    {
+      "epoch": 9.785096473719229,
+      "grad_norm": 0.2586815357208252,
+      "learning_rate": 0.00010496949621769976,
+      "loss": 0.0438,
+      "step": 3670
+    },
+    {
+      "epoch": 9.798403193612774,
+      "grad_norm": 0.004748414736241102,
+      "learning_rate": 0.00010475819158237425,
+      "loss": 0.0817,
+      "step": 3675
+    },
+    {
+      "epoch": 9.81170991350632,
+      "grad_norm": 0.19157640635967255,
+      "learning_rate": 0.00010454686565151149,
+      "loss": 0.0939,
+      "step": 3680
+    },
+    {
+      "epoch": 9.825016633399867,
+      "grad_norm": 0.3925589621067047,
+      "learning_rate": 0.00010433551937091183,
+      "loss": 0.0357,
+      "step": 3685
+    },
+    {
+      "epoch": 9.838323353293413,
+      "grad_norm": 0.6299968957901001,
+      "learning_rate": 0.00010412415368646673,
+      "loss": 0.0429,
+      "step": 3690
+    },
+    {
+      "epoch": 9.851630073186959,
+      "grad_norm": 0.003008866449818015,
+      "learning_rate": 0.00010391276954415444,
+      "loss": 0.2409,
+      "step": 3695
+    },
+    {
+      "epoch": 9.864936793080506,
+      "grad_norm": 33.798118591308594,
+      "learning_rate": 0.00010370136789003582,
+      "loss": 0.2173,
+      "step": 3700
+    },
+    {
+      "epoch": 9.878243512974052,
+      "grad_norm": 2.2869341373443604,
+      "learning_rate": 0.00010348994967025012,
+      "loss": 0.1955,
+      "step": 3705
+    },
+    {
+      "epoch": 9.891550232867598,
+      "grad_norm": 0.3005884885787964,
+      "learning_rate": 0.00010327851583101071,
+      "loss": 0.0804,
+      "step": 3710
+    },
+    {
+      "epoch": 9.904856952761145,
+      "grad_norm": 0.15614235401153564,
+      "learning_rate": 0.0001030670673186009,
+      "loss": 0.0821,
+      "step": 3715
+    },
+    {
+      "epoch": 9.918163672654691,
+      "grad_norm": 24.48946762084961,
+      "learning_rate": 0.00010285560507936961,
+      "loss": 0.0234,
+      "step": 3720
+    },
+    {
+      "epoch": 9.931470392548237,
+      "grad_norm": 0.33835986256599426,
+      "learning_rate": 0.00010264413005972735,
+      "loss": 0.0957,
+      "step": 3725
+    },
+    {
+      "epoch": 9.944777112441782,
+      "grad_norm": 64.79618835449219,
+      "learning_rate": 0.00010243264320614157,
+      "loss": 0.0583,
+      "step": 3730
+    },
+    {
+      "epoch": 9.95808383233533,
+      "grad_norm": 0.7911556959152222,
+      "learning_rate": 0.00010222114546513295,
+      "loss": 0.0557,
+      "step": 3735
+    },
+    {
+      "epoch": 9.971390552228875,
+      "grad_norm": 0.6701344847679138,
+      "learning_rate": 0.0001020096377832707,
+      "loss": 0.153,
+      "step": 3740
+    },
+    {
+      "epoch": 9.984697272122421,
+      "grad_norm": 1.700467586517334,
+      "learning_rate": 0.00010179812110716864,
+      "loss": 0.1336,
+      "step": 3745
+    },
+    {
+      "epoch": 9.998003992015969,
+      "grad_norm": 0.4652709364891052,
+      "learning_rate": 0.00010158659638348081,
+      "loss": 0.036,
+      "step": 3750
+    },
+    {
+      "epoch": 9.998003992015969,
+      "eval_loss": 1.1191763877868652,
+      "eval_macro_f1": 67.80213350209692,
+      "eval_macro_precision": 71.58253936792221,
+      "eval_macro_recall": 65.93726965528465,
+      "eval_micro_f1": 87.5374251497006,
+      "eval_micro_precision": 87.5374251497006,
+      "eval_micro_recall": 87.5374251497006,
+      "eval_runtime": 5.0299,
+      "eval_samples_per_second": 531.228,
+      "eval_steps_per_second": 33.202,
+      "step": 3750
+    },
+    {
+      "epoch": 10.013306719893546,
+      "grad_norm": 0.45541515946388245,
+      "learning_rate": 0.00010137506455889721,
+      "loss": 0.0613,
+      "step": 3755
+    },
+    {
+      "epoch": 10.026613439787093,
+      "grad_norm": 0.1207580417394638,
+      "learning_rate": 0.00010116352658013973,
+      "loss": 0.0425,
+      "step": 3760
+    },
+    {
+      "epoch": 10.039920159680639,
+      "grad_norm": 0.4036962389945984,
+      "learning_rate": 0.00010095198339395769,
+      "loss": 0.0186,
+      "step": 3765
+    },
+    {
+      "epoch": 10.053226879574185,
+      "grad_norm": 0.08803912997245789,
+      "learning_rate": 0.00010074043594712379,
+      "loss": 0.0239,
+      "step": 3770
+    },
+    {
+      "epoch": 10.066533599467732,
+      "grad_norm": 0.4955994188785553,
+      "learning_rate": 0.00010052888518642978,
+      "loss": 0.0191,
+      "step": 3775
+    },
+    {
+      "epoch": 10.079840319361278,
+      "grad_norm": 0.512108564376831,
+      "learning_rate": 0.00010031733205868224,
+      "loss": 0.0547,
+      "step": 3780
+    },
+    {
+      "epoch": 10.093147039254823,
+      "grad_norm": 0.367607057094574,
+      "learning_rate": 0.00010010577751069838,
+      "loss": 0.0353,
+      "step": 3785
+    },
+    {
+      "epoch": 10.10645375914837,
+      "grad_norm": 0.5691599249839783,
+      "learning_rate": 9.989422248930167e-05,
+      "loss": 0.0286,
+      "step": 3790
+    },
+    {
+      "epoch": 10.119760479041917,
+      "grad_norm": 0.006430297624319792,
+      "learning_rate": 9.968266794131777e-05,
+      "loss": 0.0087,
+      "step": 3795
+    },
+    {
+      "epoch": 10.133067198935462,
+      "grad_norm": 0.43629810214042664,
+      "learning_rate": 9.947111481357021e-05,
+      "loss": 0.0256,
+      "step": 3800
+    },
+    {
+      "epoch": 10.146373918829008,
+      "grad_norm": 0.5085962414741516,
+      "learning_rate": 9.925956405287624e-05,
+      "loss": 0.0159,
+      "step": 3805
+    },
+    {
+      "epoch": 10.159680638722556,
+      "grad_norm": 5.319223880767822,
+      "learning_rate": 9.904801660604234e-05,
+      "loss": 0.0566,
+      "step": 3810
+    },
+    {
+      "epoch": 10.172987358616101,
+      "grad_norm": 0.47697970271110535,
+      "learning_rate": 9.883647341986032e-05,
+      "loss": 0.1049,
+      "step": 3815
+    },
+    {
+      "epoch": 10.186294078509647,
+      "grad_norm": 25.805103302001953,
+      "learning_rate": 9.862493544110282e-05,
+      "loss": 0.2494,
+      "step": 3820
+    },
+    {
+      "epoch": 10.199600798403194,
+      "grad_norm": 0.004839545115828514,
+      "learning_rate": 9.84134036165192e-05,
+      "loss": 0.0705,
+      "step": 3825
+    },
+    {
+      "epoch": 10.21290751829674,
+      "grad_norm": 29.43570899963379,
+      "learning_rate": 9.820187889283137e-05,
+      "loss": 0.2269,
+      "step": 3830
+    },
+    {
+      "epoch": 10.226214238190286,
+      "grad_norm": 0.4234377145767212,
+      "learning_rate": 9.799036221672931e-05,
+      "loss": 0.055,
+      "step": 3835
+    },
+    {
+      "epoch": 10.239520958083832,
+      "grad_norm": 0.11629313975572586,
+      "learning_rate": 9.777885453486706e-05,
+      "loss": 0.0419,
+      "step": 3840
+    },
+    {
+      "epoch": 10.252827677977379,
+      "grad_norm": 1.7031394243240356,
+      "learning_rate": 9.756735679385844e-05,
+      "loss": 0.2843,
+      "step": 3845
+    },
+    {
+      "epoch": 10.266134397870925,
+      "grad_norm": 0.0188790000975132,
+      "learning_rate": 9.735586994027267e-05,
+      "loss": 0.0727,
+      "step": 3850
+    },
+    {
+      "epoch": 10.27944111776447,
+      "grad_norm": 0.007724573370069265,
+      "learning_rate": 9.71443949206304e-05,
+      "loss": 0.0565,
+      "step": 3855
+    },
+    {
+      "epoch": 10.292747837658018,
+      "grad_norm": 0.6192372441291809,
+      "learning_rate": 9.693293268139914e-05,
+      "loss": 0.0634,
+      "step": 3860
+    },
+    {
+      "epoch": 10.306054557551564,
+      "grad_norm": 0.005998663604259491,
+      "learning_rate": 9.672148416898932e-05,
+      "loss": 0.0553,
+      "step": 3865
+    },
+    {
+      "epoch": 10.31936127744511,
+      "grad_norm": 0.7205235362052917,
+      "learning_rate": 9.651005032974994e-05,
+      "loss": 0.012,
+      "step": 3870
+    },
+    {
+      "epoch": 10.332667997338657,
+      "grad_norm": 0.14316412806510925,
+      "learning_rate": 9.629863210996419e-05,
+      "loss": 0.0101,
+      "step": 3875
+    },
+    {
+      "epoch": 10.345974717232203,
+      "grad_norm": 0.015486049465835094,
+      "learning_rate": 9.608723045584557e-05,
+      "loss": 0.0649,
+      "step": 3880
+    },
+    {
+      "epoch": 10.359281437125748,
+      "grad_norm": 1.8815991878509521,
+      "learning_rate": 9.587584631353329e-05,
+      "loss": 0.1858,
+      "step": 3885
+    },
+    {
+      "epoch": 10.372588157019294,
+      "grad_norm": 0.5233331918716431,
+      "learning_rate": 9.566448062908819e-05,
+      "loss": 0.0641,
+      "step": 3890
+    },
+    {
+      "epoch": 10.385894876912841,
+      "grad_norm": 0.5349389910697937,
+      "learning_rate": 9.545313434848856e-05,
+      "loss": 0.0432,
+      "step": 3895
+    },
+    {
+      "epoch": 10.399201596806387,
+      "grad_norm": 0.001954582752659917,
+      "learning_rate": 9.524180841762577e-05,
+      "loss": 0.0079,
+      "step": 3900
+    },
+    {
+      "epoch": 10.412508316699933,
+      "grad_norm": 0.19347253441810608,
+      "learning_rate": 9.503050378230022e-05,
+      "loss": 0.1334,
+      "step": 3905
+    },
+    {
+      "epoch": 10.42581503659348,
+      "grad_norm": 41.07929611206055,
+      "learning_rate": 9.481922138821696e-05,
+      "loss": 0.1135,
+      "step": 3910
+    },
+    {
+      "epoch": 10.439121756487026,
+      "grad_norm": 0.6698688268661499,
+      "learning_rate": 9.460796218098143e-05,
+      "loss": 0.0256,
+      "step": 3915
+    },
+    {
+      "epoch": 10.452428476380572,
+      "grad_norm": 0.2920777201652527,
+      "learning_rate": 9.439672710609531e-05,
+      "loss": 0.0303,
+      "step": 3920
+    },
+    {
+      "epoch": 10.46573519627412,
+      "grad_norm": 0.0010179243981838226,
+      "learning_rate": 9.418551710895243e-05,
+      "loss": 0.0186,
+      "step": 3925
+    },
+    {
+      "epoch": 10.479041916167665,
+      "grad_norm": 29.91852569580078,
+      "learning_rate": 9.397433313483416e-05,
+      "loss": 0.2275,
+      "step": 3930
+    },
+    {
+      "epoch": 10.49234863606121,
+      "grad_norm": 5.829182147979736,
+      "learning_rate": 9.376317612890556e-05,
+      "loss": 0.2251,
+      "step": 3935
+    },
+    {
+      "epoch": 10.505655355954758,
+      "grad_norm": 0.5852800011634827,
+      "learning_rate": 9.355204703621093e-05,
+      "loss": 0.1962,
+      "step": 3940
+    },
+    {
+      "epoch": 10.518962075848304,
+      "grad_norm": 0.12783049046993256,
+      "learning_rate": 9.334094680166962e-05,
+      "loss": 0.0286,
+      "step": 3945
+    },
+    {
+      "epoch": 10.53226879574185,
+      "grad_norm": 0.3511544167995453,
+      "learning_rate": 9.312987637007191e-05,
+      "loss": 0.0544,
+      "step": 3950
+    },
+    {
+      "epoch": 10.545575515635395,
+      "grad_norm": 0.4987472891807556,
+      "learning_rate": 9.291883668607461e-05,
+      "loss": 0.079,
+      "step": 3955
+    },
+    {
+      "epoch": 10.558882235528943,
+      "grad_norm": 0.30509260296821594,
+      "learning_rate": 9.270782869419694e-05,
+      "loss": 0.0875,
+      "step": 3960
+    },
+    {
+      "epoch": 10.572188955422488,
+      "grad_norm": 0.721363365650177,
+      "learning_rate": 9.249685333881633e-05,
+      "loss": 0.029,
+      "step": 3965
+    },
+    {
+      "epoch": 10.585495675316034,
+      "grad_norm": 0.7444695234298706,
+      "learning_rate": 9.228591156416404e-05,
+      "loss": 0.0402,
+      "step": 3970
+    },
+    {
+      "epoch": 10.598802395209582,
+      "grad_norm": 13.542842864990234,
+      "learning_rate": 9.207500431432115e-05,
+      "loss": 0.1159,
+      "step": 3975
+    },
+    {
+      "epoch": 10.612109115103127,
+      "grad_norm": 0.2932852506637573,
+      "learning_rate": 9.186413253321418e-05,
+      "loss": 0.0927,
+      "step": 3980
+    },
+    {
+      "epoch": 10.625415834996673,
+      "grad_norm": 1.3260793685913086,
+      "learning_rate": 9.165329716461087e-05,
+      "loss": 0.0972,
+      "step": 3985
+    },
+    {
+      "epoch": 10.638722554890219,
+      "grad_norm": 0.11311087757349014,
+      "learning_rate": 9.144249915211605e-05,
+      "loss": 0.0425,
+      "step": 3990
+    },
+    {
+      "epoch": 10.652029274783766,
+      "grad_norm": 0.4719153642654419,
+      "learning_rate": 9.123173943916728e-05,
+      "loss": 0.0714,
+      "step": 3995
+    },
+    {
+      "epoch": 10.665335994677312,
+      "grad_norm": 19.49868392944336,
+      "learning_rate": 9.102101896903084e-05,
+      "loss": 0.0478,
+      "step": 4000
+    },
+    {
+      "epoch": 10.678642714570858,
+      "grad_norm": 3.4599437713623047,
+      "learning_rate": 9.081033868479727e-05,
+      "loss": 0.0556,
+      "step": 4005
+    },
+    {
+      "epoch": 10.691949434464405,
+      "grad_norm": 4.2978620529174805,
+      "learning_rate": 9.059969952937727e-05,
+      "loss": 0.1521,
+      "step": 4010
+    },
+    {
+      "epoch": 10.70525615435795,
+      "grad_norm": 0.7445971965789795,
+      "learning_rate": 9.038910244549747e-05,
+      "loss": 0.0861,
+      "step": 4015
+    },
+    {
+      "epoch": 10.718562874251496,
+      "grad_norm": 0.7097308039665222,
+      "learning_rate": 9.01785483756963e-05,
+      "loss": 0.0422,
+      "step": 4020
+    },
+    {
+      "epoch": 10.731869594145044,
+      "grad_norm": 0.5130826830863953,
+      "learning_rate": 8.99680382623195e-05,
+      "loss": 0.1419,
+      "step": 4025
+    },
+    {
+      "epoch": 10.74517631403859,
+      "grad_norm": 0.006258231122046709,
+      "learning_rate": 8.975757304751626e-05,
+      "loss": 0.0277,
+      "step": 4030
+    },
+    {
+      "epoch": 10.758483033932135,
+      "grad_norm": 1.6164250373840332,
+      "learning_rate": 8.954715367323468e-05,
+      "loss": 0.1055,
+      "step": 4035
+    },
+    {
+      "epoch": 10.771789753825683,
+      "grad_norm": 0.5906782746315002,
+      "learning_rate": 8.933678108121774e-05,
+      "loss": 0.0427,
+      "step": 4040
+    },
+    {
+      "epoch": 10.785096473719229,
+      "grad_norm": 1.2181246280670166,
+      "learning_rate": 8.912645621299918e-05,
+      "loss": 0.0399,
+      "step": 4045
+    },
+    {
+      "epoch": 10.798403193612774,
+      "grad_norm": 0.004639142192900181,
+      "learning_rate": 8.891618000989891e-05,
+      "loss": 0.234,
+      "step": 4050
+    },
+    {
+      "epoch": 10.81170991350632,
+      "grad_norm": 0.5142450332641602,
+      "learning_rate": 8.870595341301921e-05,
+      "loss": 0.092,
+      "step": 4055
+    },
+    {
+      "epoch": 10.825016633399867,
+      "grad_norm": 0.17664307355880737,
+      "learning_rate": 8.849577736324033e-05,
+      "loss": 0.1276,
+      "step": 4060
+    },
+    {
+      "epoch": 10.838323353293413,
+      "grad_norm": 0.006000434514135122,
+      "learning_rate": 8.828565280121617e-05,
+      "loss": 0.0729,
+      "step": 4065
+    },
+    {
+      "epoch": 10.851630073186959,
+      "grad_norm": 0.4471026062965393,
+      "learning_rate": 8.807558066737041e-05,
+      "loss": 0.0274,
+      "step": 4070
+    },
+    {
+      "epoch": 10.864936793080506,
+      "grad_norm": 0.03893847018480301,
+      "learning_rate": 8.786556190189182e-05,
+      "loss": 0.108,
+      "step": 4075
+    },
+    {
+      "epoch": 10.878243512974052,
+      "grad_norm": 0.004344575107097626,
+      "learning_rate": 8.765559744473053e-05,
+      "loss": 0.0387,
+      "step": 4080
+    },
+    {
+      "epoch": 10.891550232867598,
+      "grad_norm": 1.100684404373169,
+      "learning_rate": 8.744568823559356e-05,
+      "loss": 0.0963,
+      "step": 4085
+    },
+    {
+      "epoch": 10.904856952761145,
+      "grad_norm": 2.4654958248138428,
+      "learning_rate": 8.723583521394054e-05,
+      "loss": 0.0638,
+      "step": 4090
+    },
+    {
+      "epoch": 10.918163672654691,
+      "grad_norm": 0.667192280292511,
+      "learning_rate": 8.702603931897982e-05,
+      "loss": 0.0457,
+      "step": 4095
+    },
+    {
+      "epoch": 10.931470392548237,
+      "grad_norm": 0.021668490022420883,
+      "learning_rate": 8.681630148966398e-05,
+      "loss": 0.001,
+      "step": 4100
+    },
+    {
+      "epoch": 10.944777112441782,
+      "grad_norm": 0.06268395483493805,
+      "learning_rate": 8.660662266468569e-05,
+      "loss": 0.0323,
+      "step": 4105
+    },
+    {
+      "epoch": 10.95808383233533,
+      "grad_norm": 0.6293454766273499,
+      "learning_rate": 8.639700378247361e-05,
+      "loss": 0.0573,
+      "step": 4110
+    },
+    {
+      "epoch": 10.971390552228875,
+      "grad_norm": 0.6691858768463135,
+      "learning_rate": 8.618744578118805e-05,
+      "loss": 0.0407,
+      "step": 4115
+    },
+    {
+      "epoch": 10.984697272122421,
+      "grad_norm": 0.014253449626266956,
+      "learning_rate": 8.597794959871694e-05,
+      "loss": 0.0089,
+      "step": 4120
+    },
+    {
+      "epoch": 10.998003992015969,
+      "grad_norm": 0.2566983997821808,
+      "learning_rate": 8.57685161726715e-05,
+      "loss": 0.0748,
+      "step": 4125
+    },
+    {
+      "epoch": 10.998003992015969,
+      "eval_loss": 1.299912929534912,
+      "eval_macro_f1": 67.6266109032924,
+      "eval_macro_precision": 70.40109716496542,
+      "eval_macro_recall": 66.54435860863018,
+      "eval_micro_f1": 87.3502994011976,
+      "eval_micro_precision": 87.3502994011976,
+      "eval_micro_recall": 87.3502994011976,
+      "eval_runtime": 4.9792,
+      "eval_samples_per_second": 536.634,
+      "eval_steps_per_second": 33.54,
+      "step": 4125
+    },
+    {
+      "epoch": 11.013306719893546,
+      "grad_norm": 0.03808118775486946,
+      "learning_rate": 8.555914644038202e-05,
+      "loss": 0.0191,
+      "step": 4130
+    },
+    {
+      "epoch": 11.026613439787093,
+      "grad_norm": 0.5897039771080017,
+      "learning_rate": 8.534984133889376e-05,
+      "loss": 0.0351,
+      "step": 4135
+    },
+    {
+      "epoch": 11.039920159680639,
+      "grad_norm": 0.0037707712035626173,
+      "learning_rate": 8.514060180496285e-05,
+      "loss": 0.0019,
+      "step": 4140
+    },
+    {
+      "epoch": 11.053226879574185,
+      "grad_norm": 0.00108764145988971,
+      "learning_rate": 8.49314287750517e-05,
+      "loss": 0.0073,
+      "step": 4145
+    },
+    {
+      "epoch": 11.066533599467732,
+      "grad_norm": 0.5410427451133728,
+      "learning_rate": 8.47223231853253e-05,
+      "loss": 0.0322,
+      "step": 4150
+    },
+    {
+      "epoch": 11.079840319361278,
+      "grad_norm": 0.6569406390190125,
+      "learning_rate": 8.451328597164679e-05,
+      "loss": 0.045,
+      "step": 4155
+    },
+    {
+      "epoch": 11.093147039254823,
+      "grad_norm": 0.0010824741329997778,
+      "learning_rate": 8.43043180695731e-05,
+      "loss": 0.0765,
+      "step": 4160
+    },
+    {
+      "epoch": 11.10645375914837,
+      "grad_norm": 0.2378658503293991,
+      "learning_rate": 8.409542041435128e-05,
+      "loss": 0.0524,
+      "step": 4165
+    },
+    {
+      "epoch": 11.119760479041917,
+      "grad_norm": 0.3880664110183716,
+      "learning_rate": 8.38865939409136e-05,
+      "loss": 0.0621,
+      "step": 4170
+    },
+    {
+      "epoch": 11.133067198935462,
+      "grad_norm": 0.7984123826026917,
+      "learning_rate": 8.367783958387407e-05,
+      "loss": 0.0756,
+      "step": 4175
+    },
+    {
+      "epoch": 11.146373918829008,
+      "grad_norm": 0.3829672634601593,
+      "learning_rate": 8.346915827752382e-05,
+      "loss": 0.0669,
+      "step": 4180
+    },
+    {
+      "epoch": 11.159680638722556,
+      "grad_norm": 0.002959012985229492,
+      "learning_rate": 8.326055095582694e-05,
+      "loss": 0.0262,
+      "step": 4185
+    },
+    {
+      "epoch": 11.172987358616101,
+      "grad_norm": 0.45035097002983093,
+      "learning_rate": 8.305201855241659e-05,
+      "loss": 0.0263,
+      "step": 4190
+    },
+    {
+      "epoch": 11.186294078509647,
+      "grad_norm": 0.002182081574574113,
+      "learning_rate": 8.284356200059051e-05,
+      "loss": 0.044,
+      "step": 4195
+    },
+    {
+      "epoch": 11.199600798403194,
+      "grad_norm": 0.0009778572712093592,
+      "learning_rate": 8.263518223330697e-05,
+      "loss": 0.0278,
+      "step": 4200
+    },
+    {
+      "epoch": 11.21290751829674,
+      "grad_norm": 0.0005438898224383593,
+      "learning_rate": 8.242688018318063e-05,
+      "loss": 0.0462,
+      "step": 4205
+    },
+    {
+      "epoch": 11.226214238190286,
+      "grad_norm": 0.0006932623218744993,
+      "learning_rate": 8.221865678247821e-05,
+      "loss": 0.0438,
+      "step": 4210
+    },
+    {
+      "epoch": 11.239520958083832,
+      "grad_norm": 0.5347298383712769,
+      "learning_rate": 8.201051296311462e-05,
+      "loss": 0.0421,
+      "step": 4215
+    },
+    {
+      "epoch": 11.252827677977379,
+      "grad_norm": 0.5404713749885559,
+      "learning_rate": 8.180244965664845e-05,
+      "loss": 0.0566,
+      "step": 4220
+    },
+    {
+      "epoch": 11.266134397870925,
+      "grad_norm": 0.9088073968887329,
+      "learning_rate": 8.159446779427797e-05,
+      "loss": 0.0362,
+      "step": 4225
+    },
+    {
+      "epoch": 11.27944111776447,
+      "grad_norm": 0.4720342457294464,
+      "learning_rate": 8.1386568306837e-05,
+      "loss": 0.1158,
+      "step": 4230
+    },
+    {
+      "epoch": 11.292747837658018,
+      "grad_norm": 47.885772705078125,
+      "learning_rate": 8.117875212479069e-05,
+      "loss": 0.0529,
+      "step": 4235
+    },
+    {
+      "epoch": 11.306054557551564,
+      "grad_norm": 0.6843175888061523,
+      "learning_rate": 8.09710201782312e-05,
+      "loss": 0.0287,
+      "step": 4240
+    },
+    {
+      "epoch": 11.31936127744511,
+      "grad_norm": 1.8459924459457397,
+      "learning_rate": 8.076337339687394e-05,
+      "loss": 0.0635,
+      "step": 4245
+    },
+    {
+      "epoch": 11.332667997338657,
+      "grad_norm": 3.7261202335357666,
+      "learning_rate": 8.055581271005292e-05,
+      "loss": 0.0271,
+      "step": 4250
+    },
+    {
+      "epoch": 11.345974717232203,
+      "grad_norm": 0.002110959030687809,
+      "learning_rate": 8.034833904671698e-05,
+      "loss": 0.0213,
+      "step": 4255
+    },
+    {
+      "epoch": 11.359281437125748,
+      "grad_norm": 0.783439040184021,
+      "learning_rate": 8.014095333542548e-05,
+      "loss": 0.157,
+      "step": 4260
+    },
+    {
+      "epoch": 11.372588157019294,
+      "grad_norm": 0.3574387729167938,
+      "learning_rate": 7.9933656504344e-05,
+      "loss": 0.0669,
+      "step": 4265
+    },
+    {
+      "epoch": 11.385894876912841,
+      "grad_norm": 0.5715398788452148,
+      "learning_rate": 7.97264494812405e-05,
+      "loss": 0.0349,
+      "step": 4270
+    },
+    {
+      "epoch": 11.399201596806387,
+      "grad_norm": 0.6373995542526245,
+      "learning_rate": 7.951933319348095e-05,
+      "loss": 0.0257,
+      "step": 4275
+    },
+    {
+      "epoch": 11.412508316699933,
+      "grad_norm": 2.8598508834838867,
+      "learning_rate": 7.931230856802514e-05,
+      "loss": 0.1056,
+      "step": 4280
+    },
+    {
+      "epoch": 11.42581503659348,
+      "grad_norm": 0.3895857334136963,
+      "learning_rate": 7.91053765314228e-05,
+      "loss": 0.047,
+      "step": 4285
+    },
+    {
+      "epoch": 11.439121756487026,
+      "grad_norm": 0.00783681869506836,
+      "learning_rate": 7.889853800980904e-05,
+      "loss": 0.0344,
+      "step": 4290
+    },
+    {
+      "epoch": 11.452428476380572,
+      "grad_norm": 0.6813995838165283,
+      "learning_rate": 7.869179392890065e-05,
+      "loss": 0.0847,
+      "step": 4295
+    },
+    {
+      "epoch": 11.46573519627412,
+      "grad_norm": 0.03324246034026146,
+      "learning_rate": 7.848514521399166e-05,
+      "loss": 0.0171,
+      "step": 4300
+    },
+    {
+      "epoch": 11.479041916167665,
+      "grad_norm": 0.012428533285856247,
+      "learning_rate": 7.827859278994925e-05,
+      "loss": 0.0697,
+      "step": 4305
+    },
+    {
+      "epoch": 11.49234863606121,
+      "grad_norm": 0.02753024362027645,
+      "learning_rate": 7.807213758120966e-05,
+      "loss": 0.037,
+      "step": 4310
+    },
+    {
+      "epoch": 11.505655355954758,
+      "grad_norm": 0.28044065833091736,
+      "learning_rate": 7.786578051177415e-05,
+      "loss": 0.0139,
+      "step": 4315
+    },
+    {
+      "epoch": 11.518962075848304,
+      "grad_norm": 0.004422112833708525,
+      "learning_rate": 7.765952250520459e-05,
+      "loss": 0.0564,
+      "step": 4320
+    },
+    {
+      "epoch": 11.53226879574185,
+      "grad_norm": 0.5593070387840271,
+      "learning_rate": 7.745336448461959e-05,
+      "loss": 0.0538,
+      "step": 4325
+    },
+    {
+      "epoch": 11.545575515635395,
+      "grad_norm": 1.2178239822387695,
+      "learning_rate": 7.72473073726902e-05,
+      "loss": 0.0725,
+      "step": 4330
+    },
+    {
+      "epoch": 11.558882235528943,
+      "grad_norm": 0.47392886877059937,
+      "learning_rate": 7.704135209163589e-05,
+      "loss": 0.0196,
+      "step": 4335
+    },
+    {
+      "epoch": 11.572188955422488,
+      "grad_norm": 0.4012060761451721,
+      "learning_rate": 7.683549956322043e-05,
+      "loss": 0.0067,
+      "step": 4340
+    },
+    {
+      "epoch": 11.585495675316034,
+      "grad_norm": 0.3064373731613159,
+      "learning_rate": 7.662975070874761e-05,
+      "loss": 0.0347,
+      "step": 4345
+    },
+    {
+      "epoch": 11.598802395209582,
+      "grad_norm": 0.0003016276750713587,
+      "learning_rate": 7.642410644905726e-05,
+      "loss": 0.0082,
+      "step": 4350
+    },
+    {
+      "epoch": 11.612109115103127,
+      "grad_norm": 0.6510899662971497,
+      "learning_rate": 7.62185677045212e-05,
+      "loss": 0.0775,
+      "step": 4355
+    },
+    {
+      "epoch": 11.625415834996673,
+      "grad_norm": 0.44786661863327026,
+      "learning_rate": 7.601313539503879e-05,
+      "loss": 0.0881,
+      "step": 4360
+    },
+    {
+      "epoch": 11.638722554890219,
+      "grad_norm": 0.16136354207992554,
+      "learning_rate": 7.580781044003324e-05,
+      "loss": 0.0321,
+      "step": 4365
+    },
+    {
+      "epoch": 11.652029274783766,
+      "grad_norm": 0.0006294685881584883,
+      "learning_rate": 7.560259375844718e-05,
+      "loss": 0.1577,
+      "step": 4370
+    },
+    {
+      "epoch": 11.665335994677312,
+      "grad_norm": 0.0005464013665914536,
+      "learning_rate": 7.539748626873866e-05,
+      "loss": 0.0069,
+      "step": 4375
+    },
+    {
+      "epoch": 11.678642714570858,
+      "grad_norm": 0.7603722810745239,
+      "learning_rate": 7.519248888887716e-05,
+      "loss": 0.0595,
+      "step": 4380
+    },
+    {
+      "epoch": 11.691949434464405,
+      "grad_norm": 43.034149169921875,
+      "learning_rate": 7.498760253633909e-05,
+      "loss": 0.1499,
+      "step": 4385
+    },
+    {
+      "epoch": 11.70525615435795,
+      "grad_norm": 0.6510688066482544,
+      "learning_rate": 7.478282812810428e-05,
+      "loss": 0.0514,
+      "step": 4390
+    },
+    {
+      "epoch": 11.718562874251496,
+      "grad_norm": 0.516620397567749,
+      "learning_rate": 7.457816658065134e-05,
+      "loss": 0.047,
+      "step": 4395
+    },
+    {
+      "epoch": 11.731869594145044,
+      "grad_norm": 0.039791446179151535,
+      "learning_rate": 7.437361880995375e-05,
+      "loss": 0.0415,
+      "step": 4400
+    },
+    {
+      "epoch": 11.74517631403859,
+      "grad_norm": 0.0013697980903089046,
+      "learning_rate": 7.416918573147588e-05,
+      "loss": 0.0077,
+      "step": 4405
+    },
+    {
+      "epoch": 11.758483033932135,
+      "grad_norm": 0.24411720037460327,
+      "learning_rate": 7.39648682601688e-05,
+      "loss": 0.0435,
+      "step": 4410
+    },
+    {
+      "epoch": 11.771789753825683,
+      "grad_norm": 0.012387442402541637,
+      "learning_rate": 7.376066731046605e-05,
+      "loss": 0.0466,
+      "step": 4415
+    },
+    {
+      "epoch": 11.785096473719229,
+      "grad_norm": 0.5403460264205933,
+      "learning_rate": 7.35565837962798e-05,
+      "loss": 0.112,
+      "step": 4420
+    },
+    {
+      "epoch": 11.798403193612774,
+      "grad_norm": 0.0008433172479271889,
+      "learning_rate": 7.335261863099651e-05,
+      "loss": 0.0437,
+      "step": 4425
+    },
+    {
+      "epoch": 11.81170991350632,
+      "grad_norm": 0.009126711636781693,
+      "learning_rate": 7.314877272747306e-05,
+      "loss": 0.0739,
+      "step": 4430
+    },
+    {
+      "epoch": 11.825016633399867,
+      "grad_norm": 0.0014366311952471733,
+      "learning_rate": 7.294504699803257e-05,
+      "loss": 0.0525,
+      "step": 4435
+    },
+    {
+      "epoch": 11.838323353293413,
+      "grad_norm": 0.8768602609634399,
+      "learning_rate": 7.274144235446023e-05,
+      "loss": 0.0247,
+      "step": 4440
+    },
+    {
+      "epoch": 11.851630073186959,
+      "grad_norm": 0.8969868421554565,
+      "learning_rate": 7.253795970799935e-05,
+      "loss": 0.0722,
+      "step": 4445
+    },
+    {
+      "epoch": 11.864936793080506,
+      "grad_norm": 3.381582498550415,
+      "learning_rate": 7.23345999693473e-05,
+      "loss": 0.0392,
+      "step": 4450
+    },
+    {
+      "epoch": 11.878243512974052,
+      "grad_norm": 0.00464739790186286,
+      "learning_rate": 7.213136404865124e-05,
+      "loss": 0.0479,
+      "step": 4455
+    },
+    {
+      "epoch": 11.891550232867598,
+      "grad_norm": 0.4277834892272949,
+      "learning_rate": 7.192825285550431e-05,
+      "loss": 0.0691,
+      "step": 4460
+    },
+    {
+      "epoch": 11.904856952761145,
+      "grad_norm": 1.2095205783843994,
+      "learning_rate": 7.172526729894129e-05,
+      "loss": 0.0572,
+      "step": 4465
+    },
+    {
+      "epoch": 11.918163672654691,
+      "grad_norm": 0.0006847070762887597,
+      "learning_rate": 7.152240828743477e-05,
+      "loss": 0.0374,
+      "step": 4470
+    },
+    {
+      "epoch": 11.931470392548237,
+      "grad_norm": 0.4241344630718231,
+      "learning_rate": 7.131967672889101e-05,
+      "loss": 0.0469,
+      "step": 4475
+    },
+    {
+      "epoch": 11.944777112441782,
+      "grad_norm": 0.0002920299884863198,
+      "learning_rate": 7.111707353064565e-05,
+      "loss": 0.0809,
+      "step": 4480
+    },
+    {
+      "epoch": 11.95808383233533,
+      "grad_norm": 0.0005238928133621812,
+      "learning_rate": 7.09145995994601e-05,
+      "loss": 0.0426,
+      "step": 4485
+    },
+    {
+      "epoch": 11.971390552228875,
+      "grad_norm": 0.0008192436653189361,
+      "learning_rate": 7.071225584151708e-05,
+      "loss": 0.037,
+      "step": 4490
+    },
+    {
+      "epoch": 11.984697272122421,
+      "grad_norm": 0.0003719358937814832,
+      "learning_rate": 7.051004316241672e-05,
+      "loss": 0.0411,
+      "step": 4495
+    },
+    {
+      "epoch": 11.998003992015969,
+      "grad_norm": 0.4025455713272095,
+      "learning_rate": 7.030796246717255e-05,
+      "loss": 0.0644,
+      "step": 4500
+    },
+    {
+      "epoch": 11.998003992015969,
+      "eval_loss": 1.445887804031372,
+      "eval_macro_f1": 67.62150339419195,
+      "eval_macro_precision": 71.442956060202,
+      "eval_macro_recall": 65.1796170254859,
+      "eval_micro_f1": 87.5374251497006,
+      "eval_micro_precision": 87.5374251497006,
+      "eval_micro_recall": 87.5374251497006,
+      "eval_runtime": 6.051,
+      "eval_samples_per_second": 441.581,
+      "eval_steps_per_second": 27.599,
+      "step": 4500
+    },
+    {
+      "epoch": 12.013306719893546,
+      "grad_norm": 0.4702771008014679,
+      "learning_rate": 7.010601466020732e-05,
+      "loss": 0.0361,
+      "step": 4505
+    },
+    {
+      "epoch": 12.026613439787093,
+      "grad_norm": 0.32089468836784363,
+      "learning_rate": 6.990420064534914e-05,
+      "loss": 0.0383,
+      "step": 4510
+    },
+    {
+      "epoch": 12.039920159680639,
+      "grad_norm": 0.1299007534980774,
+      "learning_rate": 6.970252132582728e-05,
+      "loss": 0.061,
+      "step": 4515
+    },
+    {
+      "epoch": 12.053226879574185,
+      "grad_norm": 0.5304166078567505,
+      "learning_rate": 6.950097760426814e-05,
+      "loss": 0.0288,
+      "step": 4520
+    },
+    {
+      "epoch": 12.066533599467732,
+      "grad_norm": 0.3846903443336487,
+      "learning_rate": 6.929957038269122e-05,
+      "loss": 0.0377,
+      "step": 4525
+    },
+    {
+      "epoch": 12.079840319361278,
+      "grad_norm": 0.001077175373211503,
+      "learning_rate": 6.909830056250527e-05,
+      "loss": 0.1997,
+      "step": 4530
+    },
+    {
+      "epoch": 12.093147039254823,
+      "grad_norm": 0.31744185090065,
+      "learning_rate": 6.889716904450387e-05,
+      "loss": 0.0564,
+      "step": 4535
+    },
+    {
+      "epoch": 12.10645375914837,
+      "grad_norm": 1.2166818380355835,
+      "learning_rate": 6.869617672886182e-05,
+      "loss": 0.0387,
+      "step": 4540
+    },
+    {
+      "epoch": 12.119760479041917,
+      "grad_norm": 0.4173411428928375,
+      "learning_rate": 6.849532451513074e-05,
+      "loss": 0.0349,
+      "step": 4545
+    },
+    {
+      "epoch": 12.133067198935462,
+      "grad_norm": 1.1611005067825317,
+      "learning_rate": 6.829461330223533e-05,
+      "loss": 0.0532,
+      "step": 4550
+    },
+    {
+      "epoch": 12.146373918829008,
+      "grad_norm": 0.9703657627105713,
+      "learning_rate": 6.809404398846921e-05,
+      "loss": 0.0657,
+      "step": 4555
+    },
+    {
+      "epoch": 12.159680638722556,
+      "grad_norm": 0.5941152572631836,
+      "learning_rate": 6.789361747149093e-05,
+      "loss": 0.0221,
+      "step": 4560
+    },
+    {
+      "epoch": 12.172987358616101,
+      "grad_norm": 0.5168139934539795,
+      "learning_rate": 6.769333464831991e-05,
+      "loss": 0.0487,
+      "step": 4565
+    },
+    {
+      "epoch": 12.186294078509647,
+      "grad_norm": 0.01413114182651043,
+      "learning_rate": 6.74931964153325e-05,
+      "loss": 0.0097,
+      "step": 4570
+    },
+    {
+      "epoch": 12.199600798403194,
+      "grad_norm": 0.00606010714545846,
+      "learning_rate": 6.729320366825784e-05,
+      "loss": 0.0238,
+      "step": 4575
+    },
+    {
+      "epoch": 12.21290751829674,
+      "grad_norm": 0.9490993022918701,
+      "learning_rate": 6.709335730217412e-05,
+      "loss": 0.0697,
+      "step": 4580
+    },
+    {
+      "epoch": 12.226214238190286,
+      "grad_norm": 0.00019976045587100089,
+      "learning_rate": 6.68936582115042e-05,
+      "loss": 0.0,
+      "step": 4585
+    },
+    {
+      "epoch": 12.239520958083832,
+      "grad_norm": 0.4100409150123596,
+      "learning_rate": 6.669410729001193e-05,
+      "loss": 0.0244,
+      "step": 4590
+    },
+    {
+      "epoch": 12.252827677977379,
+      "grad_norm": 0.21931087970733643,
+      "learning_rate": 6.649470543079798e-05,
+      "loss": 0.0459,
+      "step": 4595
+    },
+    {
+      "epoch": 12.266134397870925,
+      "grad_norm": 0.39032813906669617,
+      "learning_rate": 6.629545352629582e-05,
+      "loss": 0.028,
+      "step": 4600
+    },
+    {
+      "epoch": 12.27944111776447,
+      "grad_norm": 0.2530187964439392,
+      "learning_rate": 6.609635246826794e-05,
+      "loss": 0.011,
+      "step": 4605
+    },
+    {
+      "epoch": 12.292747837658018,
+      "grad_norm": 0.000762270821724087,
+      "learning_rate": 6.589740314780157e-05,
+      "loss": 0.0347,
+      "step": 4610
+    },
+    {
+      "epoch": 12.306054557551564,
+      "grad_norm": 0.05708887055516243,
+      "learning_rate": 6.569860645530487e-05,
+      "loss": 0.0406,
+      "step": 4615
+    },
+    {
+      "epoch": 12.31936127744511,
+      "grad_norm": 0.00042223682976327837,
+      "learning_rate": 6.549996328050296e-05,
+      "loss": 0.0387,
+      "step": 4620
+    },
+    {
+      "epoch": 12.332667997338657,
+      "grad_norm": 0.4647948741912842,
+      "learning_rate": 6.530147451243377e-05,
+      "loss": 0.0285,
+      "step": 4625
+    },
+    {
+      "epoch": 12.345974717232203,
+      "grad_norm": 0.94704669713974,
+      "learning_rate": 6.51031410394443e-05,
+      "loss": 0.0515,
+      "step": 4630
+    },
+    {
+      "epoch": 12.359281437125748,
+      "grad_norm": 0.0007556902128271759,
+      "learning_rate": 6.490496374918647e-05,
+      "loss": 0.0334,
+      "step": 4635
+    },
+    {
+      "epoch": 12.372588157019294,
+      "grad_norm": 0.5485817193984985,
+      "learning_rate": 6.470694352861312e-05,
+      "loss": 0.0679,
+      "step": 4640
+    },
+    {
+      "epoch": 12.385894876912841,
+      "grad_norm": 0.6135960817337036,
+      "learning_rate": 6.450908126397423e-05,
+      "loss": 0.0247,
+      "step": 4645
+    },
+    {
+      "epoch": 12.399201596806387,
+      "grad_norm": 0.5665143728256226,
+      "learning_rate": 6.431137784081282e-05,
+      "loss": 0.0321,
+      "step": 4650
+    },
+    {
+      "epoch": 12.412508316699933,
+      "grad_norm": 0.0015934518305584788,
+      "learning_rate": 6.411383414396095e-05,
+      "loss": 0.0294,
+      "step": 4655
+    },
+    {
+      "epoch": 12.42581503659348,
+      "grad_norm": 0.3874810039997101,
+      "learning_rate": 6.391645105753583e-05,
+      "loss": 0.0295,
+      "step": 4660
+    },
+    {
+      "epoch": 12.439121756487026,
+      "grad_norm": 0.000483459240058437,
+      "learning_rate": 6.371922946493591e-05,
+      "loss": 0.0359,
+      "step": 4665
+    },
+    {
+      "epoch": 12.452428476380572,
+      "grad_norm": 0.520750105381012,
+      "learning_rate": 6.352217024883678e-05,
+      "loss": 0.0393,
+      "step": 4670
+    },
+    {
+      "epoch": 12.46573519627412,
+      "grad_norm": 0.5864303112030029,
+      "learning_rate": 6.33252742911874e-05,
+      "loss": 0.003,
+      "step": 4675
+    },
+    {
+      "epoch": 12.479041916167665,
+      "grad_norm": 1.0039427280426025,
+      "learning_rate": 6.312854247320595e-05,
+      "loss": 0.1379,
+      "step": 4680
+    },
+    {
+      "epoch": 12.49234863606121,
+      "grad_norm": 0.006187811028212309,
+      "learning_rate": 6.293197567537605e-05,
+      "loss": 0.024,
+      "step": 4685
+    },
+    {
+      "epoch": 12.505655355954758,
+      "grad_norm": 0.615149974822998,
+      "learning_rate": 6.273557477744285e-05,
+      "loss": 0.0604,
+      "step": 4690
+    },
+    {
+      "epoch": 12.518962075848304,
+      "grad_norm": 0.006822746712714434,
+      "learning_rate": 6.25393406584088e-05,
+      "loss": 0.0376,
+      "step": 4695
+    },
+    {
+      "epoch": 12.53226879574185,
+      "grad_norm": 0.038054272532463074,
+      "learning_rate": 6.234327419653013e-05,
+      "loss": 0.0287,
+      "step": 4700
+    },
+    {
+      "epoch": 12.545575515635395,
+      "grad_norm": 0.0010650916956365108,
+      "learning_rate": 6.214737626931261e-05,
+      "loss": 0.0987,
+      "step": 4705
+    },
+    {
+      "epoch": 12.558882235528943,
+      "grad_norm": 1.0322421789169312,
+      "learning_rate": 6.19516477535077e-05,
+      "loss": 0.055,
+      "step": 4710
+    },
+    {
+      "epoch": 12.572188955422488,
+      "grad_norm": 0.006024849601089954,
+      "learning_rate": 6.175608952510874e-05,
+      "loss": 0.0016,
+      "step": 4715
+    },
+    {
+      "epoch": 12.585495675316034,
+      "grad_norm": 0.0465017631649971,
+      "learning_rate": 6.156070245934684e-05,
+      "loss": 0.0503,
+      "step": 4720
+    },
+    {
+      "epoch": 12.598802395209582,
+      "grad_norm": 0.2706870436668396,
+      "learning_rate": 6.136548743068713e-05,
+      "loss": 0.0268,
+      "step": 4725
+    },
+    {
+      "epoch": 12.612109115103127,
+      "grad_norm": 0.0009190389537252486,
+      "learning_rate": 6.117044531282481e-05,
+      "loss": 0.0928,
+      "step": 4730
+    },
+    {
+      "epoch": 12.625415834996673,
+      "grad_norm": 0.6977292895317078,
+      "learning_rate": 6.097557697868108e-05,
+      "loss": 0.0543,
+      "step": 4735
+    },
+    {
+      "epoch": 12.638722554890219,
+      "grad_norm": 1.7532901763916016,
+      "learning_rate": 6.078088330039945e-05,
+      "loss": 0.0318,
+      "step": 4740
+    },
+    {
+      "epoch": 12.652029274783766,
+      "grad_norm": 0.0035180707927793264,
+      "learning_rate": 6.0586365149341806e-05,
+      "loss": 0.0384,
+      "step": 4745
+    },
+    {
+      "epoch": 12.665335994677312,
+      "grad_norm": 0.2537969648838043,
+      "learning_rate": 6.039202339608432e-05,
+      "loss": 0.0376,
+      "step": 4750
+    },
+    {
+      "epoch": 12.678642714570858,
+      "grad_norm": 0.0006269970326684415,
+      "learning_rate": 6.019785891041381e-05,
+      "loss": 0.0568,
+      "step": 4755
+    },
+    {
+      "epoch": 12.691949434464405,
+      "grad_norm": 0.91341233253479,
+      "learning_rate": 6.0003872561323584e-05,
+      "loss": 0.0295,
+      "step": 4760
+    },
+    {
+      "epoch": 12.70525615435795,
+      "grad_norm": 0.3654066324234009,
+      "learning_rate": 5.9810065217009824e-05,
+      "loss": 0.0188,
+      "step": 4765
+    },
+    {
+      "epoch": 12.718562874251496,
+      "grad_norm": 0.0023132911883294582,
+      "learning_rate": 5.9616437744867535e-05,
+      "loss": 0.0302,
+      "step": 4770
+    },
+    {
+      "epoch": 12.731869594145044,
+      "grad_norm": 0.3889198303222656,
+      "learning_rate": 5.9422991011486626e-05,
+      "loss": 0.0503,
+      "step": 4775
+    },
+    {
+      "epoch": 12.74517631403859,
+      "grad_norm": 0.0006781243137083948,
+      "learning_rate": 5.922972588264818e-05,
+      "loss": 0.0005,
+      "step": 4780
+    },
+    {
+      "epoch": 12.758483033932135,
+      "grad_norm": 0.0015274898614734411,
+      "learning_rate": 5.9036643223320475e-05,
+      "loss": 0.029,
+      "step": 4785
+    },
+    {
+      "epoch": 12.771789753825683,
+      "grad_norm": 0.000553864287212491,
+      "learning_rate": 5.8843743897655077e-05,
+      "loss": 0.0619,
+      "step": 4790
+    },
+    {
+      "epoch": 12.785096473719229,
+      "grad_norm": 0.12268496304750443,
+      "learning_rate": 5.865102876898315e-05,
+      "loss": 0.0525,
+      "step": 4795
+    },
+    {
+      "epoch": 12.798403193612774,
+      "grad_norm": 0.5278739929199219,
+      "learning_rate": 5.845849869981137e-05,
+      "loss": 0.0105,
+      "step": 4800
+    },
+    {
+      "epoch": 12.81170991350632,
+      "grad_norm": 0.3189643621444702,
+      "learning_rate": 5.8266154551818216e-05,
+      "loss": 0.051,
+      "step": 4805
+    },
+    {
+      "epoch": 12.825016633399867,
+      "grad_norm": 0.5433342456817627,
+      "learning_rate": 5.807399718585009e-05,
+      "loss": 0.042,
+      "step": 4810
+    },
+    {
+      "epoch": 12.838323353293413,
+      "grad_norm": 0.943836510181427,
+      "learning_rate": 5.788202746191734e-05,
+      "loss": 0.0255,
+      "step": 4815
+    },
+    {
+      "epoch": 12.851630073186959,
+      "grad_norm": 0.1590261310338974,
+      "learning_rate": 5.769024623919064e-05,
+      "loss": 0.0127,
+      "step": 4820
+    },
+    {
+      "epoch": 12.864936793080506,
+      "grad_norm": 0.45719876885414124,
+      "learning_rate": 5.749865437599703e-05,
+      "loss": 0.0495,
+      "step": 4825
+    },
+    {
+      "epoch": 12.878243512974052,
+      "grad_norm": 0.031068850308656693,
+      "learning_rate": 5.7307252729815833e-05,
+      "loss": 0.0578,
+      "step": 4830
+    },
+    {
+      "epoch": 12.891550232867598,
+      "grad_norm": 0.00038968288572505116,
+      "learning_rate": 5.711604215727532e-05,
+      "loss": 0.0487,
+      "step": 4835
+    },
+    {
+      "epoch": 12.904856952761145,
+      "grad_norm": 0.028543902561068535,
+      "learning_rate": 5.6925023514148414e-05,
+      "loss": 0.0064,
+      "step": 4840
+    },
+    {
+      "epoch": 12.918163672654691,
+      "grad_norm": 0.000108074709714856,
+      "learning_rate": 5.6734197655349156e-05,
+      "loss": 0.0306,
+      "step": 4845
+    },
+    {
+      "epoch": 12.931470392548237,
+      "grad_norm": 0.0013804432237520814,
+      "learning_rate": 5.654356543492882e-05,
+      "loss": 0.0457,
+      "step": 4850
+    },
+    {
+      "epoch": 12.944777112441782,
+      "grad_norm": 0.00011231685493839905,
+      "learning_rate": 5.635312770607179e-05,
+      "loss": 0.0294,
+      "step": 4855
+    },
+    {
+      "epoch": 12.95808383233533,
+      "grad_norm": 0.45072370767593384,
+      "learning_rate": 5.616288532109225e-05,
+      "loss": 0.063,
+      "step": 4860
+    },
+    {
+      "epoch": 12.971390552228875,
+      "grad_norm": 0.6160675287246704,
+      "learning_rate": 5.597283913143006e-05,
+      "loss": 0.0382,
+      "step": 4865
+    },
+    {
+      "epoch": 12.984697272122421,
+      "grad_norm": 1.4277915954589844,
+      "learning_rate": 5.5782989987646896e-05,
+      "loss": 0.0962,
+      "step": 4870
+    },
+    {
+      "epoch": 12.998003992015969,
+      "grad_norm": 0.00010680627747206017,
+      "learning_rate": 5.559333873942259e-05,
+      "loss": 0.0201,
+      "step": 4875
+    },
+    {
+      "epoch": 12.998003992015969,
+      "eval_loss": 1.546578049659729,
+      "eval_macro_f1": 67.89414685540753,
+      "eval_macro_precision": 71.53678356588046,
+      "eval_macro_recall": 65.58376331013532,
+      "eval_micro_f1": 87.6497005988024,
+      "eval_micro_precision": 87.6497005988024,
+      "eval_micro_recall": 87.6497005988024,
+      "eval_runtime": 5.2905,
+      "eval_samples_per_second": 505.058,
+      "eval_steps_per_second": 31.566,
+      "step": 4875
+    },
+    {
+      "epoch": 13.013306719893546,
+      "grad_norm": 0.6065590381622314,
+      "learning_rate": 5.5403886235551374e-05,
+      "loss": 0.0549,
+      "step": 4880
+    },
+    {
+      "epoch": 13.026613439787093,
+      "grad_norm": 0.001139428117312491,
+      "learning_rate": 5.521463332393784e-05,
+      "loss": 0.0363,
+      "step": 4885
+    },
+    {
+      "epoch": 13.039920159680639,
+      "grad_norm": 4.570841701934114e-05,
+      "learning_rate": 5.5025580851593436e-05,
+      "loss": 0.0332,
+      "step": 4890
+    },
+    {
+      "epoch": 13.053226879574185,
+      "grad_norm": 0.013228015042841434,
+      "learning_rate": 5.483672966463245e-05,
+      "loss": 0.004,
+      "step": 4895
+    },
+    {
+      "epoch": 13.066533599467732,
+      "grad_norm": 0.8052898049354553,
+      "learning_rate": 5.4648080608268245e-05,
+      "loss": 0.0337,
+      "step": 4900
+    },
+    {
+      "epoch": 13.079840319361278,
+      "grad_norm": 0.652985692024231,
+      "learning_rate": 5.445963452680973e-05,
+      "loss": 0.0274,
+      "step": 4905
+    },
+    {
+      "epoch": 13.093147039254823,
+      "grad_norm": 0.8381507992744446,
+      "learning_rate": 5.427139226365718e-05,
+      "loss": 0.0304,
+      "step": 4910
+    },
+    {
+      "epoch": 13.10645375914837,
+      "grad_norm": 0.9545891284942627,
+      "learning_rate": 5.4083354661298814e-05,
+      "loss": 0.0852,
+      "step": 4915
+    },
+    {
+      "epoch": 13.119760479041917,
+      "grad_norm": 0.7544320821762085,
+      "learning_rate": 5.38955225613069e-05,
+      "loss": 0.0799,
+      "step": 4920
+    },
+    {
+      "epoch": 13.133067198935462,
+      "grad_norm": 0.5602716207504272,
+      "learning_rate": 5.3707896804333756e-05,
+      "loss": 0.0518,
+      "step": 4925
+    },
+    {
+      "epoch": 13.146373918829008,
+      "grad_norm": 0.823800265789032,
+      "learning_rate": 5.3520478230108464e-05,
+      "loss": 0.0438,
+      "step": 4930
+    },
+    {
+      "epoch": 13.159680638722556,
+      "grad_norm": 0.0064308131113648415,
+      "learning_rate": 5.333326767743263e-05,
+      "loss": 0.0392,
+      "step": 4935
+    },
+    {
+      "epoch": 13.172987358616101,
+      "grad_norm": 1.1453149318695068,
+      "learning_rate": 5.314626598417707e-05,
+      "loss": 0.0568,
+      "step": 4940
+    },
+    {
+      "epoch": 13.186294078509647,
+      "grad_norm": 0.00013595831114798784,
+      "learning_rate": 5.295947398727763e-05,
+      "loss": 0.0475,
+      "step": 4945
+    },
+    {
+      "epoch": 13.199600798403194,
+      "grad_norm": 0.4676132798194885,
+      "learning_rate": 5.277289252273174e-05,
+      "loss": 0.0579,
+      "step": 4950
+    },
+    {
+      "epoch": 13.21290751829674,
+      "grad_norm": 0.41389578580856323,
+      "learning_rate": 5.258652242559461e-05,
+      "loss": 0.0186,
+      "step": 4955
+    },
+    {
+      "epoch": 13.226214238190286,
+      "grad_norm": 0.5707477927207947,
+      "learning_rate": 5.2400364529975446e-05,
+      "loss": 0.0515,
+      "step": 4960
+    },
+    {
+      "epoch": 13.239520958083832,
+      "grad_norm": 5.477669037645683e-05,
+      "learning_rate": 5.221441966903371e-05,
+      "loss": 0.0461,
+      "step": 4965
+    },
+    {
+      "epoch": 13.252827677977379,
+      "grad_norm": 0.34287309646606445,
+      "learning_rate": 5.2028688674975415e-05,
+      "loss": 0.0395,
+      "step": 4970
+    },
+    {
+      "epoch": 13.266134397870925,
+      "grad_norm": 6.0640715673798695e-05,
+      "learning_rate": 5.184317237904939e-05,
+      "loss": 0.0,
+      "step": 4975
+    },
+    {
+      "epoch": 13.27944111776447,
+      "grad_norm": 0.3991676867008209,
+      "learning_rate": 5.1657871611543605e-05,
+      "loss": 0.009,
+      "step": 4980
+    },
+    {
+      "epoch": 13.292747837658018,
+      "grad_norm": 5.605097976513207e-05,
+      "learning_rate": 5.147278720178148e-05,
+      "loss": 0.0523,
+      "step": 4985
+    },
+    {
+      "epoch": 13.306054557551564,
+      "grad_norm": 0.42107030749320984,
+      "learning_rate": 5.128791997811799e-05,
+      "loss": 0.0158,
+      "step": 4990
+    },
+    {
+      "epoch": 13.31936127744511,
+      "grad_norm": 0.00029585917945951223,
+      "learning_rate": 5.110327076793613e-05,
+      "loss": 0.0316,
+      "step": 4995
+    },
+    {
+      "epoch": 13.332667997338657,
+      "grad_norm": 0.6536248326301575,
+      "learning_rate": 5.091884039764321e-05,
+      "loss": 0.0242,
+      "step": 5000
+    },
+    {
+      "epoch": 13.345974717232203,
+      "grad_norm": 0.00014492319314740598,
+      "learning_rate": 5.0734629692667e-05,
+      "loss": 0.0297,
+      "step": 5005
+    },
+    {
+      "epoch": 13.359281437125748,
+      "grad_norm": 0.9105699062347412,
+      "learning_rate": 5.055063947745233e-05,
+      "loss": 0.0343,
+      "step": 5010
+    },
+    {
+      "epoch": 13.372588157019294,
+      "grad_norm": 0.00011185540643054992,
+      "learning_rate": 5.036687057545704e-05,
+      "loss": 0.0215,
+      "step": 5015
+    },
+    {
+      "epoch": 13.385894876912841,
+      "grad_norm": 0.3092271685600281,
+      "learning_rate": 5.01833238091485e-05,
+      "loss": 0.0789,
+      "step": 5020
+    },
+    {
+      "epoch": 13.399201596806387,
+      "grad_norm": 0.3386293053627014,
+      "learning_rate": 5.000000000000002e-05,
+      "loss": 0.0668,
+      "step": 5025
+    },
+    {
+      "epoch": 13.412508316699933,
+      "grad_norm": 0.001697537605650723,
+      "learning_rate": 4.98168999684869e-05,
+      "loss": 0.0337,
+      "step": 5030
+    },
+    {
+      "epoch": 13.42581503659348,
+      "grad_norm": 0.39653119444847107,
+      "learning_rate": 4.9634024534083044e-05,
+      "loss": 0.012,
+      "step": 5035
+    },
+    {
+      "epoch": 13.439121756487026,
+      "grad_norm": 0.19999262690544128,
+      "learning_rate": 4.945137451525707e-05,
+      "loss": 0.0134,
+      "step": 5040
+    },
+    {
+      "epoch": 13.452428476380572,
+      "grad_norm": 0.2290111631155014,
+      "learning_rate": 4.926895072946871e-05,
+      "loss": 0.02,
+      "step": 5045
+    },
+    {
+      "epoch": 13.46573519627412,
+      "grad_norm": 0.39010921120643616,
+      "learning_rate": 4.908675399316534e-05,
+      "loss": 0.0259,
+      "step": 5050
+    },
+    {
+      "epoch": 13.479041916167665,
+      "grad_norm": 1.3865301609039307,
+      "learning_rate": 4.890478512177795e-05,
+      "loss": 0.0423,
+      "step": 5055
+    },
+    {
+      "epoch": 13.49234863606121,
+      "grad_norm": 0.7625669240951538,
+      "learning_rate": 4.8723044929717906e-05,
+      "loss": 0.0528,
+      "step": 5060
+    },
+    {
+      "epoch": 13.505655355954758,
+      "grad_norm": 0.45569607615470886,
+      "learning_rate": 4.8541534230372974e-05,
+      "loss": 0.0291,
+      "step": 5065
+    },
+    {
+      "epoch": 13.518962075848304,
+      "grad_norm": 8.144270395860076e-05,
+      "learning_rate": 4.836025383610382e-05,
+      "loss": 0.0211,
+      "step": 5070
+    },
+    {
+      "epoch": 13.53226879574185,
+      "grad_norm": 0.4941138029098511,
+      "learning_rate": 4.8179204558240444e-05,
+      "loss": 0.076,
+      "step": 5075
+    },
+    {
+      "epoch": 13.545575515635395,
+      "grad_norm": 0.008551147766411304,
+      "learning_rate": 4.799838720707846e-05,
+      "loss": 0.0175,
+      "step": 5080
+    },
+    {
+      "epoch": 13.558882235528943,
+      "grad_norm": 0.4467279314994812,
+      "learning_rate": 4.7817802591875426e-05,
+      "loss": 0.0756,
+      "step": 5085
+    },
+    {
+      "epoch": 13.572188955422488,
+      "grad_norm": 0.6588059067726135,
+      "learning_rate": 4.76374515208473e-05,
+      "loss": 0.0569,
+      "step": 5090
+    },
+    {
+      "epoch": 13.585495675316034,
+      "grad_norm": 0.3777833580970764,
+      "learning_rate": 4.7457334801164775e-05,
+      "loss": 0.0071,
+      "step": 5095
+    },
+    {
+      "epoch": 13.598802395209582,
+      "grad_norm": 0.0004773762484546751,
+      "learning_rate": 4.727745323894976e-05,
+      "loss": 0.0472,
+      "step": 5100
+    },
+    {
+      "epoch": 13.612109115103127,
+      "grad_norm": 0.41729772090911865,
+      "learning_rate": 4.7097807639271683e-05,
+      "loss": 0.049,
+      "step": 5105
+    },
+    {
+      "epoch": 13.625415834996673,
+      "grad_norm": 0.00019701718701981008,
+      "learning_rate": 4.691839880614389e-05,
+      "loss": 0.0167,
+      "step": 5110
+    },
+    {
+      "epoch": 13.638722554890219,
+      "grad_norm": 0.540219247341156,
+      "learning_rate": 4.673922754252002e-05,
+      "loss": 0.0383,
+      "step": 5115
+    },
+    {
+      "epoch": 13.652029274783766,
+      "grad_norm": 0.1812252551317215,
+      "learning_rate": 4.656029465029057e-05,
+      "loss": 0.04,
+      "step": 5120
+    },
+    {
+      "epoch": 13.665335994677312,
+      "grad_norm": 0.8494847416877747,
+      "learning_rate": 4.638160093027908e-05,
+      "loss": 0.039,
+      "step": 5125
+    },
+    {
+      "epoch": 13.678642714570858,
+      "grad_norm": 0.5483289957046509,
+      "learning_rate": 4.620314718223876e-05,
+      "loss": 0.0531,
+      "step": 5130
+    },
+    {
+      "epoch": 13.691949434464405,
+      "grad_norm": 1.038064956665039,
+      "learning_rate": 4.6024934204848745e-05,
+      "loss": 0.0554,
+      "step": 5135
+    },
+    {
+      "epoch": 13.70525615435795,
+      "grad_norm": 9.937622962752357e-05,
+      "learning_rate": 4.5846962795710556e-05,
+      "loss": 0.0337,
+      "step": 5140
+    },
+    {
+      "epoch": 13.718562874251496,
+      "grad_norm": 1.4000746011734009,
+      "learning_rate": 4.566923375134472e-05,
+      "loss": 0.04,
+      "step": 5145
+    },
+    {
+      "epoch": 13.731869594145044,
+      "grad_norm": 0.0001771347306203097,
+      "learning_rate": 4.549174786718684e-05,
+      "loss": 0.0263,
+      "step": 5150
+    },
+    {
+      "epoch": 13.74517631403859,
+      "grad_norm": 0.5845703482627869,
+      "learning_rate": 4.5314505937584417e-05,
+      "loss": 0.0491,
+      "step": 5155
+    },
+    {
+      "epoch": 13.758483033932135,
+      "grad_norm": 0.49106454849243164,
+      "learning_rate": 4.513750875579303e-05,
+      "loss": 0.1028,
+      "step": 5160
+    },
+    {
+      "epoch": 13.771789753825683,
+      "grad_norm": 0.4913535416126251,
+      "learning_rate": 4.4960757113972864e-05,
+      "loss": 0.0562,
+      "step": 5165
+    },
+    {
+      "epoch": 13.785096473719229,
+      "grad_norm": 0.6593743562698364,
+      "learning_rate": 4.4784251803185226e-05,
+      "loss": 0.0398,
+      "step": 5170
+    },
+    {
+      "epoch": 13.798403193612774,
+      "grad_norm": 0.0005124734016135335,
+      "learning_rate": 4.4607993613388976e-05,
+      "loss": 0.027,
+      "step": 5175
+    },
+    {
+      "epoch": 13.81170991350632,
+      "grad_norm": 0.0022483225911855698,
+      "learning_rate": 4.44319833334369e-05,
+      "loss": 0.0327,
+      "step": 5180
+    },
+    {
+      "epoch": 13.825016633399867,
+      "grad_norm": 0.0013437450397759676,
+      "learning_rate": 4.425622175107229e-05,
+      "loss": 0.0602,
+      "step": 5185
+    },
+    {
+      "epoch": 13.838323353293413,
+      "grad_norm": 0.46366244554519653,
+      "learning_rate": 4.4080709652925336e-05,
+      "loss": 0.0184,
+      "step": 5190
+    },
+    {
+      "epoch": 13.851630073186959,
+      "grad_norm": 0.00015787413576617837,
+      "learning_rate": 4.390544782450971e-05,
+      "loss": 0.072,
+      "step": 5195
+    },
+    {
+      "epoch": 13.864936793080506,
+      "grad_norm": 0.00038810219848528504,
+      "learning_rate": 4.373043705021899e-05,
+      "loss": 0.0315,
+      "step": 5200
+    },
+    {
+      "epoch": 13.878243512974052,
+      "grad_norm": 0.00165260536596179,
+      "learning_rate": 4.355567811332311e-05,
+      "loss": 0.0077,
+      "step": 5205
+    },
+    {
+      "epoch": 13.891550232867598,
+      "grad_norm": 0.6691223382949829,
+      "learning_rate": 4.338117179596485e-05,
+      "loss": 0.0221,
+      "step": 5210
+    },
+    {
+      "epoch": 13.904856952761145,
+      "grad_norm": 0.0013225431321188807,
+      "learning_rate": 4.32069188791565e-05,
+      "loss": 0.0422,
+      "step": 5215
+    },
+    {
+      "epoch": 13.918163672654691,
+      "grad_norm": 0.0003881505399476737,
+      "learning_rate": 4.3032920142776125e-05,
+      "loss": 0.0415,
+      "step": 5220
+    },
+    {
+      "epoch": 13.931470392548237,
+      "grad_norm": 0.46118324995040894,
+      "learning_rate": 4.285917636556429e-05,
+      "loss": 0.0262,
+      "step": 5225
+    },
+    {
+      "epoch": 13.944777112441782,
+      "grad_norm": 0.0037599815987050533,
+      "learning_rate": 4.26856883251204e-05,
+      "loss": 0.0271,
+      "step": 5230
+    },
+    {
+      "epoch": 13.95808383233533,
+      "grad_norm": 0.5800197720527649,
+      "learning_rate": 4.251245679789928e-05,
+      "loss": 0.0481,
+      "step": 5235
+    },
+    {
+      "epoch": 13.971390552228875,
+      "grad_norm": 0.5036775469779968,
+      "learning_rate": 4.233948255920785e-05,
+      "loss": 0.0188,
+      "step": 5240
+    },
+    {
+      "epoch": 13.984697272122421,
+      "grad_norm": 0.5240902900695801,
+      "learning_rate": 4.216676638320135e-05,
+      "loss": 0.0113,
+      "step": 5245
+    },
+    {
+      "epoch": 13.998003992015969,
+      "grad_norm": 0.0005705656949430704,
+      "learning_rate": 4.19943090428802e-05,
+      "loss": 0.01,
+      "step": 5250
+    },
+    {
+      "epoch": 13.998003992015969,
+      "eval_loss": 1.5540084838867188,
+      "eval_macro_f1": 68.44116182144285,
+      "eval_macro_precision": 70.37403623594724,
+      "eval_macro_recall": 67.0800347172364,
+      "eval_micro_f1": 87.3877245508982,
+      "eval_micro_precision": 87.3877245508982,
+      "eval_micro_recall": 87.3877245508982,
+      "eval_runtime": 4.7976,
+      "eval_samples_per_second": 556.948,
+      "eval_steps_per_second": 34.809,
+      "step": 5250
+    },
+    {
+      "epoch": 14.013306719893546,
+      "grad_norm": 0.44576719403266907,
+      "learning_rate": 4.182211131008628e-05,
+      "loss": 0.0421,
+      "step": 5255
+    },
+    {
+      "epoch": 14.026613439787093,
+      "grad_norm": 0.0001893198787001893,
+      "learning_rate": 4.1650173955499585e-05,
+      "loss": 0.053,
+      "step": 5260
+    },
+    {
+      "epoch": 14.039920159680639,
+      "grad_norm": 0.00020121457055211067,
+      "learning_rate": 4.147849774863488e-05,
+      "loss": 0.0323,
+      "step": 5265
+    },
+    {
+      "epoch": 14.053226879574185,
+      "grad_norm": 0.37343332171440125,
+      "learning_rate": 4.1307083457838e-05,
+      "loss": 0.0444,
+      "step": 5270
+    },
+    {
+      "epoch": 14.066533599467732,
+      "grad_norm": 0.0002352256269659847,
+      "learning_rate": 4.1135931850282726e-05,
+      "loss": 0.0324,
+      "step": 5275
+    },
+    {
+      "epoch": 14.079840319361278,
+      "grad_norm": 0.4024161398410797,
+      "learning_rate": 4.096504369196704e-05,
+      "loss": 0.0149,
+      "step": 5280
+    },
+    {
+      "epoch": 14.093147039254823,
+      "grad_norm": 0.6625308394432068,
+      "learning_rate": 4.0794419747709886e-05,
+      "loss": 0.012,
+      "step": 5285
+    },
+    {
+      "epoch": 14.10645375914837,
+      "grad_norm": 0.4134993255138397,
+      "learning_rate": 4.062406078114776e-05,
+      "loss": 0.0451,
+      "step": 5290
+    },
+    {
+      "epoch": 14.119760479041917,
+      "grad_norm": 0.600199282169342,
+      "learning_rate": 4.045396755473121e-05,
+      "loss": 0.0326,
+      "step": 5295
+    },
+    {
+      "epoch": 14.133067198935462,
+      "grad_norm": 0.7752711772918701,
+      "learning_rate": 4.028414082972141e-05,
+      "loss": 0.0541,
+      "step": 5300
+    },
+    {
+      "epoch": 14.146373918829008,
+      "grad_norm": 0.0010675047524273396,
+      "learning_rate": 4.011458136618681e-05,
+      "loss": 0.0139,
+      "step": 5305
+    },
+    {
+      "epoch": 14.159680638722556,
+      "grad_norm": 0.6033095717430115,
+      "learning_rate": 3.994528992299971e-05,
+      "loss": 0.026,
+      "step": 5310
+    },
+    {
+      "epoch": 14.172987358616101,
+      "grad_norm": 0.6383993625640869,
+      "learning_rate": 3.977626725783291e-05,
+      "loss": 0.057,
+      "step": 5315
+    },
+    {
+      "epoch": 14.186294078509647,
+      "grad_norm": 0.008502921089529991,
+      "learning_rate": 3.960751412715629e-05,
+      "loss": 0.0275,
+      "step": 5320
+    },
+    {
+      "epoch": 14.199600798403194,
+      "grad_norm": 0.5096533894538879,
+      "learning_rate": 3.943903128623335e-05,
+      "loss": 0.0325,
+      "step": 5325
+    },
+    {
+      "epoch": 14.21290751829674,
+      "grad_norm": 0.00016405718633905053,
+      "learning_rate": 3.9270819489117904e-05,
+      "loss": 0.019,
+      "step": 5330
+    },
+    {
+      "epoch": 14.226214238190286,
+      "grad_norm": 0.0003083451883867383,
+      "learning_rate": 3.9102879488650757e-05,
+      "loss": 0.0253,
+      "step": 5335
+    },
+    {
+      "epoch": 14.239520958083832,
+      "grad_norm": 0.5092978477478027,
+      "learning_rate": 3.893521203645618e-05,
+      "loss": 0.0634,
+      "step": 5340
+    },
+    {
+      "epoch": 14.252827677977379,
+      "grad_norm": 0.0014424554537981749,
+      "learning_rate": 3.876781788293876e-05,
+      "loss": 0.0181,
+      "step": 5345
+    },
+    {
+      "epoch": 14.266134397870925,
+      "grad_norm": 0.0012855530949309468,
+      "learning_rate": 3.860069777727983e-05,
+      "loss": 0.0605,
+      "step": 5350
+    },
+    {
+      "epoch": 14.27944111776447,
+      "grad_norm": 0.37602004408836365,
+      "learning_rate": 3.843385246743417e-05,
+      "loss": 0.0319,
+      "step": 5355
+    },
+    {
+      "epoch": 14.292747837658018,
+      "grad_norm": 0.21032342314720154,
+      "learning_rate": 3.826728270012686e-05,
+      "loss": 0.0396,
+      "step": 5360
+    },
+    {
+      "epoch": 14.306054557551564,
+      "grad_norm": 0.36076977849006653,
+      "learning_rate": 3.810098922084958e-05,
+      "loss": 0.0277,
+      "step": 5365
+    },
+    {
+      "epoch": 14.31936127744511,
+      "grad_norm": 0.2929077446460724,
+      "learning_rate": 3.7934972773857634e-05,
+      "loss": 0.0274,
+      "step": 5370
+    },
+    {
+      "epoch": 14.332667997338657,
+      "grad_norm": 0.760208249092102,
+      "learning_rate": 3.776923410216636e-05,
+      "loss": 0.1027,
+      "step": 5375
+    },
+    {
+      "epoch": 14.345974717232203,
+      "grad_norm": 0.8322664499282837,
+      "learning_rate": 3.7603773947547874e-05,
+      "loss": 0.0629,
+      "step": 5380
+    },
+    {
+      "epoch": 14.359281437125748,
+      "grad_norm": 0.6654631495475769,
+      "learning_rate": 3.7438593050527845e-05,
+      "loss": 0.0398,
+      "step": 5385
+    },
+    {
+      "epoch": 14.372588157019294,
+      "grad_norm": 5.7775789173319936e-05,
+      "learning_rate": 3.7273692150382135e-05,
+      "loss": 0.0,
+      "step": 5390
+    },
+    {
+      "epoch": 14.385894876912841,
+      "grad_norm": 0.5650693774223328,
+      "learning_rate": 3.710907198513337e-05,
+      "loss": 0.0329,
+      "step": 5395
+    },
+    {
+      "epoch": 14.399201596806387,
+      "grad_norm": 0.00029549148166552186,
+      "learning_rate": 3.694473329154778e-05,
+      "loss": 0.0329,
+      "step": 5400
+    },
+    {
+      "epoch": 14.412508316699933,
+      "grad_norm": 0.008520903065800667,
+      "learning_rate": 3.678067680513182e-05,
+      "loss": 0.0216,
+      "step": 5405
+    },
+    {
+      "epoch": 14.42581503659348,
+      "grad_norm": 0.0003043456526938826,
+      "learning_rate": 3.661690326012897e-05,
+      "loss": 0.0083,
+      "step": 5410
+    },
+    {
+      "epoch": 14.439121756487026,
+      "grad_norm": 0.00047451278078369796,
+      "learning_rate": 3.645341338951639e-05,
+      "loss": 0.0026,
+      "step": 5415
+    },
+    {
+      "epoch": 14.452428476380572,
+      "grad_norm": 0.00206236750818789,
+      "learning_rate": 3.6290207925001584e-05,
+      "loss": 0.0085,
+      "step": 5420
+    },
+    {
+      "epoch": 14.46573519627412,
+      "grad_norm": 0.36209532618522644,
+      "learning_rate": 3.6127287597019186e-05,
+      "loss": 0.0063,
+      "step": 5425
+    },
+    {
+      "epoch": 14.479041916167665,
+      "grad_norm": 0.12500470876693726,
+      "learning_rate": 3.5964653134727776e-05,
+      "loss": 0.0282,
+      "step": 5430
+    },
+    {
+      "epoch": 14.49234863606121,
+      "grad_norm": 0.47154736518859863,
+      "learning_rate": 3.580230526600639e-05,
+      "loss": 0.0214,
+      "step": 5435
+    },
+    {
+      "epoch": 14.505655355954758,
+      "grad_norm": 0.038162682205438614,
+      "learning_rate": 3.564024471745154e-05,
+      "loss": 0.0106,
+      "step": 5440
+    },
+    {
+      "epoch": 14.518962075848304,
+      "grad_norm": 8.110934868454933e-05,
+      "learning_rate": 3.547847221437372e-05,
+      "loss": 0.0286,
+      "step": 5445
+    },
+    {
+      "epoch": 14.53226879574185,
+      "grad_norm": 0.0009972963016480207,
+      "learning_rate": 3.531698848079425e-05,
+      "loss": 0.0303,
+      "step": 5450
+    },
+    {
+      "epoch": 14.545575515635395,
+      "grad_norm": 0.5373179912567139,
+      "learning_rate": 3.5155794239442184e-05,
+      "loss": 0.0664,
+      "step": 5455
+    },
+    {
+      "epoch": 14.558882235528943,
+      "grad_norm": 0.6623542904853821,
+      "learning_rate": 3.4994890211750754e-05,
+      "loss": 0.0555,
+      "step": 5460
+    },
+    {
+      "epoch": 14.572188955422488,
+      "grad_norm": 0.7875820398330688,
+      "learning_rate": 3.483427711785449e-05,
+      "loss": 0.0798,
+      "step": 5465
+    },
+    {
+      "epoch": 14.585495675316034,
+      "grad_norm": 0.5920644998550415,
+      "learning_rate": 3.467395567658573e-05,
+      "loss": 0.0409,
+      "step": 5470
+    },
+    {
+      "epoch": 14.598802395209582,
+      "grad_norm": 0.0001885549136204645,
+      "learning_rate": 3.45139266054715e-05,
+      "loss": 0.0487,
+      "step": 5475
+    },
+    {
+      "epoch": 14.612109115103127,
+      "grad_norm": 0.32315418124198914,
+      "learning_rate": 3.4354190620730406e-05,
+      "loss": 0.0674,
+      "step": 5480
+    },
+    {
+      "epoch": 14.625415834996673,
+      "grad_norm": 5.668059384333901e-05,
+      "learning_rate": 3.419474843726921e-05,
+      "loss": 0.0275,
+      "step": 5485
+    },
+    {
+      "epoch": 14.638722554890219,
+      "grad_norm": 0.0033752620220184326,
+      "learning_rate": 3.4035600768679855e-05,
+      "loss": 0.0251,
+      "step": 5490
+    },
+    {
+      "epoch": 14.652029274783766,
+      "grad_norm": 0.6533347964286804,
+      "learning_rate": 3.387674832723611e-05,
+      "loss": 0.0937,
+      "step": 5495
+    },
+    {
+      "epoch": 14.665335994677312,
+      "grad_norm": 0.0007474491721950471,
+      "learning_rate": 3.37181918238904e-05,
+      "loss": 0.0246,
+      "step": 5500
+    },
+    {
+      "epoch": 14.678642714570858,
+      "grad_norm": 0.6492227911949158,
+      "learning_rate": 3.3559931968270753e-05,
+      "loss": 0.0379,
+      "step": 5505
+    },
+    {
+      "epoch": 14.691949434464405,
+      "grad_norm": 0.0007334169349633157,
+      "learning_rate": 3.340196946867753e-05,
+      "loss": 0.0349,
+      "step": 5510
+    },
+    {
+      "epoch": 14.70525615435795,
+      "grad_norm": 0.00012730961316265166,
+      "learning_rate": 3.3244305032080183e-05,
+      "loss": 0.0225,
+      "step": 5515
+    },
+    {
+      "epoch": 14.718562874251496,
+      "grad_norm": 0.00010851504339370877,
+      "learning_rate": 3.308693936411421e-05,
+      "loss": 0.0278,
+      "step": 5520
+    },
+    {
+      "epoch": 14.731869594145044,
+      "grad_norm": 8.108102338155732e-05,
+      "learning_rate": 3.292987316907792e-05,
+      "loss": 0.0298,
+      "step": 5525
+    },
+    {
+      "epoch": 14.74517631403859,
+      "grad_norm": 1.1605405807495117,
+      "learning_rate": 3.2773107149929384e-05,
+      "loss": 0.1119,
+      "step": 5530
+    },
+    {
+      "epoch": 14.758483033932135,
+      "grad_norm": 0.00032222739537246525,
+      "learning_rate": 3.2616642008283213e-05,
+      "loss": 0.0169,
+      "step": 5535
+    },
+    {
+      "epoch": 14.771789753825683,
+      "grad_norm": 0.0024410784244537354,
+      "learning_rate": 3.2460478444407374e-05,
+      "loss": 0.024,
+      "step": 5540
+    },
+    {
+      "epoch": 14.785096473719229,
+      "grad_norm": 0.028958754613995552,
+      "learning_rate": 3.230461715722007e-05,
+      "loss": 0.0064,
+      "step": 5545
+    },
+    {
+      "epoch": 14.798403193612774,
+      "grad_norm": 0.7407286763191223,
+      "learning_rate": 3.21490588442868e-05,
+      "loss": 0.0552,
+      "step": 5550
+    },
+    {
+      "epoch": 14.81170991350632,
+      "grad_norm": 0.49551257491111755,
+      "learning_rate": 3.19938042018169e-05,
+      "loss": 0.0356,
+      "step": 5555
+    },
+    {
+      "epoch": 14.825016633399867,
+      "grad_norm": 0.7646135687828064,
+      "learning_rate": 3.1838853924660795e-05,
+      "loss": 0.0379,
+      "step": 5560
+    },
+    {
+      "epoch": 14.838323353293413,
+      "grad_norm": 0.000533775077201426,
+      "learning_rate": 3.1684208706306574e-05,
+      "loss": 0.0099,
+      "step": 5565
+    },
+    {
+      "epoch": 14.851630073186959,
+      "grad_norm": 0.414894163608551,
+      "learning_rate": 3.152986923887703e-05,
+      "loss": 0.0079,
+      "step": 5570
+    },
+    {
+      "epoch": 14.864936793080506,
+      "grad_norm": 0.7916562557220459,
+      "learning_rate": 3.137583621312665e-05,
+      "loss": 0.0791,
+      "step": 5575
+    },
+    {
+      "epoch": 14.878243512974052,
+      "grad_norm": 0.4727895259857178,
+      "learning_rate": 3.1222110318438304e-05,
+      "loss": 0.0614,
+      "step": 5580
+    },
+    {
+      "epoch": 14.891550232867598,
+      "grad_norm": 1.6393200159072876,
+      "learning_rate": 3.1068692242820386e-05,
+      "loss": 0.0374,
+      "step": 5585
+    },
+    {
+      "epoch": 14.904856952761145,
+      "grad_norm": 0.47209104895591736,
+      "learning_rate": 3.0915582672903556e-05,
+      "loss": 0.0177,
+      "step": 5590
+    },
+    {
+      "epoch": 14.918163672654691,
+      "grad_norm": 0.45247048139572144,
+      "learning_rate": 3.076278229393773e-05,
+      "loss": 0.0435,
+      "step": 5595
+    },
+    {
+      "epoch": 14.931470392548237,
+      "grad_norm": 1.2875940799713135,
+      "learning_rate": 3.0610291789789095e-05,
+      "loss": 0.0685,
+      "step": 5600
+    },
+    {
+      "epoch": 14.944777112441782,
+      "grad_norm": 0.000212906816159375,
+      "learning_rate": 3.0458111842936952e-05,
+      "loss": 0.0407,
+      "step": 5605
+    },
+    {
+      "epoch": 14.95808383233533,
+      "grad_norm": 0.4576803147792816,
+      "learning_rate": 3.030624313447067e-05,
+      "loss": 0.025,
+      "step": 5610
+    },
+    {
+      "epoch": 14.971390552228875,
+      "grad_norm": 0.6176482439041138,
+      "learning_rate": 3.0154686344086636e-05,
+      "loss": 0.0568,
+      "step": 5615
+    },
+    {
+      "epoch": 14.984697272122421,
+      "grad_norm": 8.765784878050908e-05,
+      "learning_rate": 3.0003442150085236e-05,
+      "loss": 0.0316,
+      "step": 5620
+    },
+    {
+      "epoch": 14.998003992015969,
+      "grad_norm": 0.6814188361167908,
+      "learning_rate": 2.9852511229367865e-05,
+      "loss": 0.0439,
+      "step": 5625
+    },
+    {
+      "epoch": 14.998003992015969,
+      "eval_loss": 1.587638258934021,
+      "eval_macro_f1": 68.64532527567134,
+      "eval_macro_precision": 70.73194829515054,
+      "eval_macro_recall": 67.08167804230526,
+      "eval_micro_f1": 87.57485029940119,
+      "eval_micro_precision": 87.57485029940119,
+      "eval_micro_recall": 87.57485029940119,
+      "eval_runtime": 4.6604,
+      "eval_samples_per_second": 573.342,
+      "eval_steps_per_second": 35.834,
+      "step": 5625
+    },
+    {
+      "epoch": 15.013306719893546,
+      "grad_norm": 0.00012128758680773899,
+      "learning_rate": 2.9701894257433826e-05,
+      "loss": 0.0248,
+      "step": 5630
+    },
+    {
+      "epoch": 15.026613439787093,
+      "grad_norm": 0.0005043774144724011,
+      "learning_rate": 2.9551591908377308e-05,
+      "loss": 0.0379,
+      "step": 5635
+    },
+    {
+      "epoch": 15.039920159680639,
+      "grad_norm": 0.00016329926438629627,
+      "learning_rate": 2.9401604854884357e-05,
+      "loss": 0.0092,
+      "step": 5640
+    },
+    {
+      "epoch": 15.053226879574185,
+      "grad_norm": 0.39029860496520996,
+      "learning_rate": 2.925193376822999e-05,
+      "loss": 0.0385,
+      "step": 5645
+    },
+    {
+      "epoch": 15.066533599467732,
+      "grad_norm": 0.40662872791290283,
+      "learning_rate": 2.9102579318274992e-05,
+      "loss": 0.0364,
+      "step": 5650
+    },
+    {
+      "epoch": 15.079840319361278,
+      "grad_norm": 0.4595581293106079,
+      "learning_rate": 2.8953542173463133e-05,
+      "loss": 0.0221,
+      "step": 5655
+    },
+    {
+      "epoch": 15.093147039254823,
+      "grad_norm": 0.5248783230781555,
+      "learning_rate": 2.880482300081797e-05,
+      "loss": 0.0464,
+      "step": 5660
+    },
+    {
+      "epoch": 15.10645375914837,
+      "grad_norm": 0.5380111336708069,
+      "learning_rate": 2.8656422465939993e-05,
+      "loss": 0.0204,
+      "step": 5665
+    },
+    {
+      "epoch": 15.119760479041917,
+      "grad_norm": 0.00021512297098524868,
+      "learning_rate": 2.8508341233003654e-05,
+      "loss": 0.0201,
+      "step": 5670
+    },
+    {
+      "epoch": 15.133067198935462,
+      "grad_norm": 0.00022878289746586233,
+      "learning_rate": 2.8360579964754274e-05,
+      "loss": 0.034,
+      "step": 5675
+    },
+    {
+      "epoch": 15.146373918829008,
+      "grad_norm": 0.0010837200097739697,
+      "learning_rate": 2.8213139322505243e-05,
+      "loss": 0.0711,
+      "step": 5680
+    },
+    {
+      "epoch": 15.159680638722556,
+      "grad_norm": 0.0003766873269341886,
+      "learning_rate": 2.8066019966134904e-05,
+      "loss": 0.0668,
+      "step": 5685
+    },
+    {
+      "epoch": 15.172987358616101,
+      "grad_norm": 0.778472900390625,
+      "learning_rate": 2.7919222554083667e-05,
+      "loss": 0.0266,
+      "step": 5690
+    },
+    {
+      "epoch": 15.186294078509647,
+      "grad_norm": 0.6452155709266663,
+      "learning_rate": 2.7772747743351135e-05,
+      "loss": 0.0261,
+      "step": 5695
+    },
+    {
+      "epoch": 15.199600798403194,
+      "grad_norm": 0.5438355803489685,
+      "learning_rate": 2.7626596189492983e-05,
+      "loss": 0.0402,
+      "step": 5700
+    },
+    {
+      "epoch": 15.21290751829674,
+      "grad_norm": 0.0052736373618245125,
+      "learning_rate": 2.7480768546618264e-05,
+      "loss": 0.0318,
+      "step": 5705
+    },
+    {
+      "epoch": 15.226214238190286,
+      "grad_norm": 0.0006670206203125417,
+      "learning_rate": 2.733526546738624e-05,
+      "loss": 0.0063,
+      "step": 5710
+    },
+    {
+      "epoch": 15.239520958083832,
+      "grad_norm": 0.00027919738204218447,
+      "learning_rate": 2.719008760300359e-05,
+      "loss": 0.0122,
+      "step": 5715
+    },
+    {
+      "epoch": 15.252827677977379,
+      "grad_norm": 0.0005136179970577359,
+      "learning_rate": 2.704523560322153e-05,
+      "loss": 0.0162,
+      "step": 5720
+    },
+    {
+      "epoch": 15.266134397870925,
+      "grad_norm": 0.003726151306182146,
+      "learning_rate": 2.690071011633284e-05,
+      "loss": 0.0202,
+      "step": 5725
+    },
+    {
+      "epoch": 15.27944111776447,
+      "grad_norm": 0.9570088982582092,
+      "learning_rate": 2.6756511789168925e-05,
+      "loss": 0.0452,
+      "step": 5730
+    },
+    {
+      "epoch": 15.292747837658018,
+      "grad_norm": 0.5644978284835815,
+      "learning_rate": 2.6612641267097005e-05,
+      "loss": 0.0692,
+      "step": 5735
+    },
+    {
+      "epoch": 15.306054557551564,
+      "grad_norm": 0.6403547525405884,
+      "learning_rate": 2.6469099194017143e-05,
+      "loss": 0.0542,
+      "step": 5740
+    },
+    {
+      "epoch": 15.31936127744511,
+      "grad_norm": 0.001306506572291255,
+      "learning_rate": 2.6325886212359498e-05,
+      "loss": 0.0193,
+      "step": 5745
+    },
+    {
+      "epoch": 15.332667997338657,
+      "grad_norm": 0.5799793601036072,
+      "learning_rate": 2.618300296308135e-05,
+      "loss": 0.0315,
+      "step": 5750
+    },
+    {
+      "epoch": 15.345974717232203,
+      "grad_norm": 0.0005930495681241155,
+      "learning_rate": 2.6040450085664158e-05,
+      "loss": 0.0418,
+      "step": 5755
+    },
+    {
+      "epoch": 15.359281437125748,
+      "grad_norm": 0.6801490783691406,
+      "learning_rate": 2.589822821811083e-05,
+      "loss": 0.0348,
+      "step": 5760
+    },
+    {
+      "epoch": 15.372588157019294,
+      "grad_norm": 0.5044436454772949,
+      "learning_rate": 2.5756337996942892e-05,
+      "loss": 0.0147,
+      "step": 5765
+    },
+    {
+      "epoch": 15.385894876912841,
+      "grad_norm": 0.5001690983772278,
+      "learning_rate": 2.561478005719743e-05,
+      "loss": 0.0757,
+      "step": 5770
+    },
+    {
+      "epoch": 15.399201596806387,
+      "grad_norm": 0.42732563614845276,
+      "learning_rate": 2.5473555032424533e-05,
+      "loss": 0.0433,
+      "step": 5775
+    },
+    {
+      "epoch": 15.412508316699933,
+      "grad_norm": 0.5017266869544983,
+      "learning_rate": 2.5332663554684222e-05,
+      "loss": 0.0308,
+      "step": 5780
+    },
+    {
+      "epoch": 15.42581503659348,
+      "grad_norm": 7.58219975978136e-05,
+      "learning_rate": 2.519210625454369e-05,
+      "loss": 0.0003,
+      "step": 5785
+    },
+    {
+      "epoch": 15.439121756487026,
+      "grad_norm": 0.00032350822584703565,
+      "learning_rate": 2.5051883761074614e-05,
+      "loss": 0.0352,
+      "step": 5790
+    },
+    {
+      "epoch": 15.452428476380572,
+      "grad_norm": 0.5080671906471252,
+      "learning_rate": 2.491199670185008e-05,
+      "loss": 0.019,
+      "step": 5795
+    },
+    {
+      "epoch": 15.46573519627412,
+      "grad_norm": 0.6078925132751465,
+      "learning_rate": 2.477244570294206e-05,
+      "loss": 0.0204,
+      "step": 5800
+    },
+    {
+      "epoch": 15.479041916167665,
+      "grad_norm": 0.44966697692871094,
+      "learning_rate": 2.4633231388918378e-05,
+      "loss": 0.0394,
+      "step": 5805
+    },
+    {
+      "epoch": 15.49234863606121,
+      "grad_norm": 0.0005682120099663734,
+      "learning_rate": 2.449435438284e-05,
+      "loss": 0.0241,
+      "step": 5810
+    },
+    {
+      "epoch": 15.505655355954758,
+      "grad_norm": 0.5577021837234497,
+      "learning_rate": 2.4355815306258354e-05,
+      "loss": 0.0319,
+      "step": 5815
+    },
+    {
+      "epoch": 15.518962075848304,
+      "grad_norm": 0.5054609179496765,
+      "learning_rate": 2.4217614779212315e-05,
+      "loss": 0.0302,
+      "step": 5820
+    },
+    {
+      "epoch": 15.53226879574185,
+      "grad_norm": 0.4755954444408417,
+      "learning_rate": 2.4079753420225694e-05,
+      "loss": 0.0302,
+      "step": 5825
+    },
+    {
+      "epoch": 15.545575515635395,
+      "grad_norm": 1.0765711069107056,
+      "learning_rate": 2.394223184630422e-05,
+      "loss": 0.0662,
+      "step": 5830
+    },
+    {
+      "epoch": 15.558882235528943,
+      "grad_norm": 0.550977349281311,
+      "learning_rate": 2.3805050672932928e-05,
+      "loss": 0.0355,
+      "step": 5835
+    },
+    {
+      "epoch": 15.572188955422488,
+      "grad_norm": 0.5157077312469482,
+      "learning_rate": 2.366821051407342e-05,
+      "loss": 0.0552,
+      "step": 5840
+    },
+    {
+      "epoch": 15.585495675316034,
+      "grad_norm": 0.8127491474151611,
+      "learning_rate": 2.3531711982161066e-05,
+      "loss": 0.0612,
+      "step": 5845
+    },
+    {
+      "epoch": 15.598802395209582,
+      "grad_norm": 0.0020727987866848707,
+      "learning_rate": 2.339555568810221e-05,
+      "loss": 0.0252,
+      "step": 5850
+    },
+    {
+      "epoch": 15.612109115103127,
+      "grad_norm": 0.00020015117479488254,
+      "learning_rate": 2.32597422412715e-05,
+      "loss": 0.0351,
+      "step": 5855
+    },
+    {
+      "epoch": 15.625415834996673,
+      "grad_norm": 0.859960675239563,
+      "learning_rate": 2.3124272249509226e-05,
+      "loss": 0.0501,
+      "step": 5860
+    },
+    {
+      "epoch": 15.638722554890219,
+      "grad_norm": 0.5178987979888916,
+      "learning_rate": 2.2989146319118425e-05,
+      "loss": 0.0377,
+      "step": 5865
+    },
+    {
+      "epoch": 15.652029274783766,
+      "grad_norm": 0.4624478220939636,
+      "learning_rate": 2.2854365054862382e-05,
+      "loss": 0.0519,
+      "step": 5870
+    },
+    {
+      "epoch": 15.665335994677312,
+      "grad_norm": 0.5773726105690002,
+      "learning_rate": 2.2719929059961698e-05,
+      "loss": 0.0177,
+      "step": 5875
+    },
+    {
+      "epoch": 15.678642714570858,
+      "grad_norm": 0.5505730509757996,
+      "learning_rate": 2.2585838936091754e-05,
+      "loss": 0.0285,
+      "step": 5880
+    },
+    {
+      "epoch": 15.691949434464405,
+      "grad_norm": 0.9743322730064392,
+      "learning_rate": 2.2452095283380003e-05,
+      "loss": 0.0165,
+      "step": 5885
+    },
+    {
+      "epoch": 15.70525615435795,
+      "grad_norm": 0.46247196197509766,
+      "learning_rate": 2.2318698700403174e-05,
+      "loss": 0.0538,
+      "step": 5890
+    },
+    {
+      "epoch": 15.718562874251496,
+      "grad_norm": 0.40537363290786743,
+      "learning_rate": 2.2185649784184746e-05,
+      "loss": 0.0436,
+      "step": 5895
+    },
+    {
+      "epoch": 15.731869594145044,
+      "grad_norm": 0.00572928274050355,
+      "learning_rate": 2.2052949130192136e-05,
+      "loss": 0.0434,
+      "step": 5900
+    },
+    {
+      "epoch": 15.74517631403859,
+      "grad_norm": 0.3674880862236023,
+      "learning_rate": 2.192059733233408e-05,
+      "loss": 0.0391,
+      "step": 5905
+    },
+    {
+      "epoch": 15.758483033932135,
+      "grad_norm": 0.4856221079826355,
+      "learning_rate": 2.178859498295809e-05,
+      "loss": 0.0211,
+      "step": 5910
+    },
+    {
+      "epoch": 15.771789753825683,
+      "grad_norm": 0.00038009166019037366,
+      "learning_rate": 2.1656942672847568e-05,
+      "loss": 0.0469,
+      "step": 5915
+    },
+    {
+      "epoch": 15.785096473719229,
+      "grad_norm": 0.14428001642227173,
+      "learning_rate": 2.152564099121944e-05,
+      "loss": 0.0424,
+      "step": 5920
+    },
+    {
+      "epoch": 15.798403193612774,
+      "grad_norm": 8.240526949521154e-05,
+      "learning_rate": 2.139469052572127e-05,
+      "loss": 0.0093,
+      "step": 5925
+    },
+    {
+      "epoch": 15.81170991350632,
+      "grad_norm": 0.5241426825523376,
+      "learning_rate": 2.1264091862428737e-05,
+      "loss": 0.0186,
+      "step": 5930
+    },
+    {
+      "epoch": 15.825016633399867,
+      "grad_norm": 0.45298877358436584,
+      "learning_rate": 2.113384558584307e-05,
+      "loss": 0.0947,
+      "step": 5935
+    },
+    {
+      "epoch": 15.838323353293413,
+      "grad_norm": 0.5497854948043823,
+      "learning_rate": 2.1003952278888382e-05,
+      "loss": 0.0561,
+      "step": 5940
+    },
+    {
+      "epoch": 15.851630073186959,
+      "grad_norm": 0.0001928492565639317,
+      "learning_rate": 2.087441252290897e-05,
+      "loss": 0.0088,
+      "step": 5945
+    },
+    {
+      "epoch": 15.864936793080506,
+      "grad_norm": 0.46321219205856323,
+      "learning_rate": 2.074522689766686e-05,
+      "loss": 0.0348,
+      "step": 5950
+    },
+    {
+      "epoch": 15.878243512974052,
+      "grad_norm": 0.4391571283340454,
+      "learning_rate": 2.0616395981339075e-05,
+      "loss": 0.0336,
+      "step": 5955
+    },
+    {
+      "epoch": 15.891550232867598,
+      "grad_norm": 0.5629188418388367,
+      "learning_rate": 2.0487920350515212e-05,
+      "loss": 0.0299,
+      "step": 5960
+    },
+    {
+      "epoch": 15.904856952761145,
+      "grad_norm": 0.00014769968402106315,
+      "learning_rate": 2.0359800580194764e-05,
+      "loss": 0.0195,
+      "step": 5965
+    },
+    {
+      "epoch": 15.918163672654691,
+      "grad_norm": 0.00027674203738570213,
+      "learning_rate": 2.0232037243784475e-05,
+      "loss": 0.0076,
+      "step": 5970
+    },
+    {
+      "epoch": 15.931470392548237,
+      "grad_norm": 0.4338863790035248,
+      "learning_rate": 2.010463091309587e-05,
+      "loss": 0.0457,
+      "step": 5975
+    },
+    {
+      "epoch": 15.944777112441782,
+      "grad_norm": 0.7306026220321655,
+      "learning_rate": 1.9977582158342754e-05,
+      "loss": 0.0369,
+      "step": 5980
+    },
+    {
+      "epoch": 15.95808383233533,
+      "grad_norm": 0.731495201587677,
+      "learning_rate": 1.985089154813846e-05,
+      "loss": 0.0672,
+      "step": 5985
+    },
+    {
+      "epoch": 15.971390552228875,
+      "grad_norm": 8.390431321458891e-05,
+      "learning_rate": 1.9724559649493567e-05,
+      "loss": 0.0107,
+      "step": 5990
+    },
+    {
+      "epoch": 15.984697272122421,
+      "grad_norm": 0.00010008139361161739,
+      "learning_rate": 1.95985870278131e-05,
+      "loss": 0.0474,
+      "step": 5995
+    },
+    {
+      "epoch": 15.998003992015969,
+      "grad_norm": 0.9425669312477112,
+      "learning_rate": 1.947297424689414e-05,
+      "loss": 0.0628,
+      "step": 6000
+    },
+    {
+      "epoch": 15.998003992015969,
+      "eval_loss": 1.6210857629776,
+      "eval_macro_f1": 67.81923557787161,
+      "eval_macro_precision": 70.7270600967691,
+      "eval_macro_recall": 65.89991990784617,
+      "eval_micro_f1": 87.4625748502994,
+      "eval_micro_precision": 87.4625748502994,
+      "eval_micro_recall": 87.4625748502994,
+      "eval_runtime": 5.0132,
+      "eval_samples_per_second": 532.989,
+      "eval_steps_per_second": 33.312,
+      "step": 6000
+    },
+    {
+      "epoch": 16.013306719893546,
+      "grad_norm": 7.936586189316586e-05,
+      "learning_rate": 1.9347721868923374e-05,
+      "loss": 0.043,
+      "step": 6005
+    },
+    {
+      "epoch": 16.02661343978709,
+      "grad_norm": 0.6429644227027893,
+      "learning_rate": 1.922283045447436e-05,
+      "loss": 0.0282,
+      "step": 6010
+    },
+    {
+      "epoch": 16.039920159680637,
+      "grad_norm": 0.0010867537930607796,
+      "learning_rate": 1.9098300562505266e-05,
+      "loss": 0.0,
+      "step": 6015
+    },
+    {
+      "epoch": 16.053226879574186,
+      "grad_norm": 0.000851222372148186,
+      "learning_rate": 1.8974132750356153e-05,
+      "loss": 0.0348,
+      "step": 6020
+    },
+    {
+      "epoch": 16.066533599467732,
+      "grad_norm": 0.0010033567668870091,
+      "learning_rate": 1.8850327573746585e-05,
+      "loss": 0.0124,
+      "step": 6025
+    },
+    {
+      "epoch": 16.079840319361278,
+      "grad_norm": 0.5936465859413147,
+      "learning_rate": 1.8726885586773212e-05,
+      "loss": 0.0336,
+      "step": 6030
+    },
+    {
+      "epoch": 16.093147039254823,
+      "grad_norm": 0.0009507182403467596,
+      "learning_rate": 1.8603807341907098e-05,
+      "loss": 0.0458,
+      "step": 6035
+    },
+    {
+      "epoch": 16.10645375914837,
+      "grad_norm": 0.42788538336753845,
+      "learning_rate": 1.8481093389991468e-05,
+      "loss": 0.0257,
+      "step": 6040
+    },
+    {
+      "epoch": 16.119760479041915,
+      "grad_norm": 0.004346577450633049,
+      "learning_rate": 1.835874428023905e-05,
+      "loss": 0.023,
+      "step": 6045
+    },
+    {
+      "epoch": 16.133067198935464,
+      "grad_norm": 5.493039861903526e-05,
+      "learning_rate": 1.8236760560229714e-05,
+      "loss": 0.0208,
+      "step": 6050
+    },
+    {
+      "epoch": 16.14637391882901,
+      "grad_norm": 4.223493306199089e-05,
+      "learning_rate": 1.8115142775908045e-05,
+      "loss": 0.0375,
+      "step": 6055
+    },
+    {
+      "epoch": 16.159680638722556,
+      "grad_norm": 0.0001787821383913979,
+      "learning_rate": 1.7993891471580893e-05,
+      "loss": 0.0135,
+      "step": 6060
+    },
+    {
+      "epoch": 16.1729873586161,
+      "grad_norm": 0.5778005123138428,
+      "learning_rate": 1.7873007189914815e-05,
+      "loss": 0.0554,
+      "step": 6065
+    },
+    {
+      "epoch": 16.186294078509647,
+      "grad_norm": 3.0017821700312197e-05,
+      "learning_rate": 1.775249047193377e-05,
+      "loss": 0.0419,
+      "step": 6070
+    },
+    {
+      "epoch": 16.199600798403193,
+      "grad_norm": 0.6397530436515808,
+      "learning_rate": 1.763234185701673e-05,
+      "loss": 0.0409,
+      "step": 6075
+    },
+    {
+      "epoch": 16.21290751829674,
+      "grad_norm": 1.0274336338043213,
+      "learning_rate": 1.7512561882895108e-05,
+      "loss": 0.0525,
+      "step": 6080
+    },
+    {
+      "epoch": 16.226214238190288,
+      "grad_norm": 0.34178104996681213,
+      "learning_rate": 1.739315108565053e-05,
+      "loss": 0.0264,
+      "step": 6085
+    },
+    {
+      "epoch": 16.239520958083833,
+      "grad_norm": 5.178775609238073e-05,
+      "learning_rate": 1.7274109999712295e-05,
+      "loss": 0.0202,
+      "step": 6090
+    },
+    {
+      "epoch": 16.25282767797738,
+      "grad_norm": 0.4585123062133789,
+      "learning_rate": 1.7155439157855037e-05,
+      "loss": 0.0302,
+      "step": 6095
+    },
+    {
+      "epoch": 16.266134397870925,
+      "grad_norm": 0.0011848661815747619,
+      "learning_rate": 1.70371390911964e-05,
+      "loss": 0.0435,
+      "step": 6100
+    },
+    {
+      "epoch": 16.27944111776447,
+      "grad_norm": 4.1747145587578416e-05,
+      "learning_rate": 1.6919210329194533e-05,
+      "loss": 0.0183,
+      "step": 6105
+    },
+    {
+      "epoch": 16.292747837658016,
+      "grad_norm": 5.5612043070141226e-05,
+      "learning_rate": 1.6801653399645866e-05,
+      "loss": 0.0358,
+      "step": 6110
+    },
+    {
+      "epoch": 16.306054557551562,
+      "grad_norm": 8.079227700363845e-05,
+      "learning_rate": 1.668446882868262e-05,
+      "loss": 0.0414,
+      "step": 6115
+    },
+    {
+      "epoch": 16.31936127744511,
+      "grad_norm": 4.4177184463478625e-05,
+      "learning_rate": 1.6567657140770475e-05,
+      "loss": 0.0208,
+      "step": 6120
+    },
+    {
+      "epoch": 16.332667997338657,
+      "grad_norm": 4.0801765862852335e-05,
+      "learning_rate": 1.6451218858706374e-05,
+      "loss": 0.0323,
+      "step": 6125
+    },
+    {
+      "epoch": 16.345974717232203,
+      "grad_norm": 0.5345805287361145,
+      "learning_rate": 1.63351545036159e-05,
+      "loss": 0.0371,
+      "step": 6130
+    },
+    {
+      "epoch": 16.35928143712575,
+      "grad_norm": 2.5667854060884565e-05,
+      "learning_rate": 1.621946459495127e-05,
+      "loss": 0.01,
+      "step": 6135
+    },
+    {
+      "epoch": 16.372588157019294,
+      "grad_norm": 6.979352474445477e-05,
+      "learning_rate": 1.610414965048874e-05,
+      "loss": 0.0508,
+      "step": 6140
+    },
+    {
+      "epoch": 16.38589487691284,
+      "grad_norm": 0.6311120986938477,
+      "learning_rate": 1.5989210186326388e-05,
+      "loss": 0.0584,
+      "step": 6145
+    },
+    {
+      "epoch": 16.39920159680639,
+      "grad_norm": 0.00010010002733906731,
+      "learning_rate": 1.587464671688187e-05,
+      "loss": 0.0332,
+      "step": 6150
+    },
+    {
+      "epoch": 16.412508316699935,
+      "grad_norm": 1.4140156507492065,
+      "learning_rate": 1.5760459754890068e-05,
+      "loss": 0.0725,
+      "step": 6155
+    },
+    {
+      "epoch": 16.42581503659348,
+      "grad_norm": 0.00011153733794344589,
+      "learning_rate": 1.5646649811400705e-05,
+      "loss": 0.0245,
+      "step": 6160
+    },
+    {
+      "epoch": 16.439121756487026,
+      "grad_norm": 0.5203162431716919,
+      "learning_rate": 1.553321739577619e-05,
+      "loss": 0.0657,
+      "step": 6165
+    },
+    {
+      "epoch": 16.45242847638057,
+      "grad_norm": 0.7140324711799622,
+      "learning_rate": 1.542016301568926e-05,
+      "loss": 0.0483,
+      "step": 6170
+    },
+    {
+      "epoch": 16.465735196274117,
+      "grad_norm": 0.6292963027954102,
+      "learning_rate": 1.5307487177120772e-05,
+      "loss": 0.0601,
+      "step": 6175
+    },
+    {
+      "epoch": 16.479041916167663,
+      "grad_norm": 3.154838486807421e-05,
+      "learning_rate": 1.5195190384357404e-05,
+      "loss": 0.0624,
+      "step": 6180
+    },
+    {
+      "epoch": 16.492348636061212,
+      "grad_norm": 0.001199195859953761,
+      "learning_rate": 1.5083273139989352e-05,
+      "loss": 0.0111,
+      "step": 6185
+    },
+    {
+      "epoch": 16.505655355954758,
+      "grad_norm": 5.260263060336001e-05,
+      "learning_rate": 1.4971735944908106e-05,
+      "loss": 0.0191,
+      "step": 6190
+    },
+    {
+      "epoch": 16.518962075848304,
+      "grad_norm": 0.5854066610336304,
+      "learning_rate": 1.4860579298304312e-05,
+      "loss": 0.0253,
+      "step": 6195
+    },
+    {
+      "epoch": 16.53226879574185,
+      "grad_norm": 0.0001449223782401532,
+      "learning_rate": 1.4749803697665366e-05,
+      "loss": 0.0163,
+      "step": 6200
+    },
+    {
+      "epoch": 16.545575515635395,
+      "grad_norm": 0.5943820476531982,
+      "learning_rate": 1.463940963877335e-05,
+      "loss": 0.0181,
+      "step": 6205
+    },
+    {
+      "epoch": 16.55888223552894,
+      "grad_norm": 0.5208426713943481,
+      "learning_rate": 1.4529397615702656e-05,
+      "loss": 0.0566,
+      "step": 6210
+    },
+    {
+      "epoch": 16.57218895542249,
+      "grad_norm": 4.5326713006943464e-05,
+      "learning_rate": 1.4419768120817889e-05,
+      "loss": 0.0406,
+      "step": 6215
+    },
+    {
+      "epoch": 16.585495675316036,
+      "grad_norm": 1.1962686777114868,
+      "learning_rate": 1.4310521644771658e-05,
+      "loss": 0.0561,
+      "step": 6220
+    },
+    {
+      "epoch": 16.59880239520958,
+      "grad_norm": 5.617663555312902e-05,
+      "learning_rate": 1.4201658676502294e-05,
+      "loss": 0.043,
+      "step": 6225
+    },
+    {
+      "epoch": 16.612109115103127,
+      "grad_norm": 0.8843726515769958,
+      "learning_rate": 1.4093179703231784e-05,
+      "loss": 0.0501,
+      "step": 6230
+    },
+    {
+      "epoch": 16.625415834996673,
+      "grad_norm": 0.5637995600700378,
+      "learning_rate": 1.3985085210463477e-05,
+      "loss": 0.0111,
+      "step": 6235
+    },
+    {
+      "epoch": 16.63872255489022,
+      "grad_norm": 0.00020396114268805832,
+      "learning_rate": 1.3877375681979943e-05,
+      "loss": 0.0233,
+      "step": 6240
+    },
+    {
+      "epoch": 16.652029274783764,
+      "grad_norm": 0.14145420491695404,
+      "learning_rate": 1.3770051599840905e-05,
+      "loss": 0.0026,
+      "step": 6245
+    },
+    {
+      "epoch": 16.665335994677314,
+      "grad_norm": 0.6490848660469055,
+      "learning_rate": 1.3663113444380905e-05,
+      "loss": 0.0318,
+      "step": 6250
+    },
+    {
+      "epoch": 16.67864271457086,
+      "grad_norm": 5.598918869509362e-05,
+      "learning_rate": 1.3556561694207338e-05,
+      "loss": 0.0113,
+      "step": 6255
+    },
+    {
+      "epoch": 16.691949434464405,
+      "grad_norm": 0.5942937731742859,
+      "learning_rate": 1.3450396826198142e-05,
+      "loss": 0.0185,
+      "step": 6260
+    },
+    {
+      "epoch": 16.70525615435795,
+      "grad_norm": 3.837359690805897e-05,
+      "learning_rate": 1.3344619315499774e-05,
+      "loss": 0.0172,
+      "step": 6265
+    },
+    {
+      "epoch": 16.718562874251496,
+      "grad_norm": 5.086205783300102e-05,
+      "learning_rate": 1.3239229635525074e-05,
+      "loss": 0.0171,
+      "step": 6270
+    },
+    {
+      "epoch": 16.731869594145042,
+      "grad_norm": 0.0009596463642083108,
+      "learning_rate": 1.3134228257951142e-05,
+      "loss": 0.0107,
+      "step": 6275
+    },
+    {
+      "epoch": 16.745176314038588,
+      "grad_norm": 0.807654619216919,
+      "learning_rate": 1.302961565271713e-05,
+      "loss": 0.0477,
+      "step": 6280
+    },
+    {
+      "epoch": 16.758483033932137,
+      "grad_norm": 0.7362765073776245,
+      "learning_rate": 1.2925392288022298e-05,
+      "loss": 0.0635,
+      "step": 6285
+    },
+    {
+      "epoch": 16.771789753825683,
+      "grad_norm": 3.490053495625034e-05,
+      "learning_rate": 1.2821558630323772e-05,
+      "loss": 0.0254,
+      "step": 6290
+    },
+    {
+      "epoch": 16.78509647371923,
+      "grad_norm": 4.203675780445337e-05,
+      "learning_rate": 1.2718115144334574e-05,
+      "loss": 0.0474,
+      "step": 6295
+    },
+    {
+      "epoch": 16.798403193612774,
+      "grad_norm": 0.419802188873291,
+      "learning_rate": 1.2615062293021507e-05,
+      "loss": 0.0356,
+      "step": 6300
+    },
+    {
+      "epoch": 16.81170991350632,
+      "grad_norm": 0.6541265249252319,
+      "learning_rate": 1.251240053760302e-05,
+      "loss": 0.0469,
+      "step": 6305
+    },
+    {
+      "epoch": 16.825016633399866,
+      "grad_norm": 0.5618201494216919,
+      "learning_rate": 1.2410130337547177e-05,
+      "loss": 0.0295,
+      "step": 6310
+    },
+    {
+      "epoch": 16.83832335329341,
+      "grad_norm": 0.6275516152381897,
+      "learning_rate": 1.230825215056971e-05,
+      "loss": 0.0205,
+      "step": 6315
+    },
+    {
+      "epoch": 16.85163007318696,
+      "grad_norm": 0.6804540157318115,
+      "learning_rate": 1.2206766432631766e-05,
+      "loss": 0.0408,
+      "step": 6320
+    },
+    {
+      "epoch": 16.864936793080506,
+      "grad_norm": 0.5253809094429016,
+      "learning_rate": 1.2105673637938053e-05,
+      "loss": 0.0787,
+      "step": 6325
+    },
+    {
+      "epoch": 16.878243512974052,
+      "grad_norm": 0.45566126704216003,
+      "learning_rate": 1.2004974218934695e-05,
+      "loss": 0.0512,
+      "step": 6330
+    },
+    {
+      "epoch": 16.891550232867598,
+      "grad_norm": 0.46057701110839844,
+      "learning_rate": 1.1904668626307226e-05,
+      "loss": 0.0786,
+      "step": 6335
+    },
+    {
+      "epoch": 16.904856952761143,
+      "grad_norm": 6.161674536997452e-05,
+      "learning_rate": 1.1804757308978654e-05,
+      "loss": 0.0475,
+      "step": 6340
+    },
+    {
+      "epoch": 16.91816367265469,
+      "grad_norm": 0.00021025777095928788,
+      "learning_rate": 1.1705240714107302e-05,
+      "loss": 0.0294,
+      "step": 6345
+    },
+    {
+      "epoch": 16.93147039254824,
+      "grad_norm": 0.00022641375835519284,
+      "learning_rate": 1.1606119287084983e-05,
+      "loss": 0.0178,
+      "step": 6350
+    },
+    {
+      "epoch": 16.944777112441784,
+      "grad_norm": 9.136064181802794e-05,
+      "learning_rate": 1.1507393471534833e-05,
+      "loss": 0.0586,
+      "step": 6355
+    },
+    {
+      "epoch": 16.95808383233533,
+      "grad_norm": 0.0001223825238412246,
+      "learning_rate": 1.1409063709309442e-05,
+      "loss": 0.0377,
+      "step": 6360
+    },
+    {
+      "epoch": 16.971390552228875,
+      "grad_norm": 0.8817633986473083,
+      "learning_rate": 1.1311130440488848e-05,
+      "loss": 0.0462,
+      "step": 6365
+    },
+    {
+      "epoch": 16.98469727212242,
+      "grad_norm": 0.43441712856292725,
+      "learning_rate": 1.1213594103378588e-05,
+      "loss": 0.0405,
+      "step": 6370
+    },
+    {
+      "epoch": 16.998003992015967,
+      "grad_norm": 0.0001357852015644312,
+      "learning_rate": 1.1116455134507664e-05,
+      "loss": 0.0,
+      "step": 6375
+    },
+    {
+      "epoch": 16.998003992015967,
+      "eval_loss": 1.636362910270691,
+      "eval_macro_f1": 68.57243488427568,
+      "eval_macro_precision": 70.72150848759064,
+      "eval_macro_recall": 66.973441952334,
+      "eval_micro_f1": 87.57485029940119,
+      "eval_micro_precision": 87.57485029940119,
+      "eval_micro_recall": 87.57485029940119,
+      "eval_runtime": 4.8106,
+      "eval_samples_per_second": 555.437,
+      "eval_steps_per_second": 34.715,
+      "step": 6375
+    },
+    {
+      "epoch": 17.013306719893546,
+      "grad_norm": 0.001735365716740489,
+      "learning_rate": 1.1019713968626632e-05,
+      "loss": 0.0342,
+      "step": 6380
+    },
+    {
+      "epoch": 17.02661343978709,
+      "grad_norm": 0.0005165593465790153,
+      "learning_rate": 1.0923371038705677e-05,
+      "loss": 0.0,
+      "step": 6385
+    },
+    {
+      "epoch": 17.039920159680637,
+      "grad_norm": 0.00010086892143590376,
+      "learning_rate": 1.0827426775932658e-05,
+      "loss": 0.0599,
+      "step": 6390
+    },
+    {
+      "epoch": 17.053226879574186,
+      "grad_norm": 7.885311060817912e-05,
+      "learning_rate": 1.0731881609711247e-05,
+      "loss": 0.0247,
+      "step": 6395
+    },
+    {
+      "epoch": 17.066533599467732,
+      "grad_norm": 0.5551512241363525,
+      "learning_rate": 1.0636735967658784e-05,
+      "loss": 0.0672,
+      "step": 6400
+    },
+    {
+      "epoch": 17.079840319361278,
+      "grad_norm": 5.913793938816525e-05,
+      "learning_rate": 1.054199027560463e-05,
+      "loss": 0.0333,
+      "step": 6405
+    },
+    {
+      "epoch": 17.093147039254823,
+      "grad_norm": 0.0009562379564158618,
+      "learning_rate": 1.0447644957588165e-05,
+      "loss": 0.0277,
+      "step": 6410
+    },
+    {
+      "epoch": 17.10645375914837,
+      "grad_norm": 0.8013795018196106,
+      "learning_rate": 1.0353700435856772e-05,
+      "loss": 0.0465,
+      "step": 6415
+    },
+    {
+      "epoch": 17.119760479041915,
+      "grad_norm": 1.1875650882720947,
+      "learning_rate": 1.026015713086418e-05,
+      "loss": 0.0421,
+      "step": 6420
+    },
+    {
+      "epoch": 17.133067198935464,
+      "grad_norm": 3.995191218564287e-05,
+      "learning_rate": 1.0167015461268304e-05,
+      "loss": 0.0602,
+      "step": 6425
+    },
+    {
+      "epoch": 17.14637391882901,
+      "grad_norm": 0.00024099642178043723,
+      "learning_rate": 1.0074275843929626e-05,
+      "loss": 0.0086,
+      "step": 6430
+    },
+    {
+      "epoch": 17.159680638722556,
+      "grad_norm": 0.9614741206169128,
+      "learning_rate": 9.98193869390922e-06,
+      "loss": 0.0321,
+      "step": 6435
+    },
+    {
+      "epoch": 17.1729873586161,
+      "grad_norm": 0.531378984451294,
+      "learning_rate": 9.890004424466825e-06,
+      "loss": 0.0268,
+      "step": 6440
+    },
+    {
+      "epoch": 17.186294078509647,
+      "grad_norm": 0.5041787028312683,
+      "learning_rate": 9.798473447059154e-06,
+      "loss": 0.0498,
+      "step": 6445
+    },
+    {
+      "epoch": 17.199600798403193,
+      "grad_norm": 0.0004790777456946671,
+      "learning_rate": 9.707346171337894e-06,
+      "loss": 0.0184,
+      "step": 6450
+    },
+    {
+      "epoch": 17.21290751829674,
+      "grad_norm": 0.00019066996173933148,
+      "learning_rate": 9.616623005147951e-06,
+      "loss": 0.0333,
+      "step": 6455
+    },
+    {
+      "epoch": 17.226214238190288,
+      "grad_norm": 5.8852790971286595e-05,
+      "learning_rate": 9.526304354525672e-06,
+      "loss": 0.008,
+      "step": 6460
+    },
+    {
+      "epoch": 17.239520958083833,
+      "grad_norm": 1.407489538192749,
+      "learning_rate": 9.436390623696911e-06,
+      "loss": 0.0333,
+      "step": 6465
+    },
+    {
+      "epoch": 17.25282767797738,
+      "grad_norm": 0.6663693189620972,
+      "learning_rate": 9.346882215075348e-06,
+      "loss": 0.0217,
+      "step": 6470
+    },
+    {
+      "epoch": 17.266134397870925,
+      "grad_norm": 0.749689519405365,
+      "learning_rate": 9.257779529260557e-06,
+      "loss": 0.0431,
+      "step": 6475
+    },
+    {
+      "epoch": 17.27944111776447,
+      "grad_norm": 0.00026045890990644693,
+      "learning_rate": 9.16908296503628e-06,
+      "loss": 0.0085,
+      "step": 6480
+    },
+    {
+      "epoch": 17.292747837658016,
+      "grad_norm": 0.8062569499015808,
+      "learning_rate": 9.080792919368696e-06,
+      "loss": 0.044,
+      "step": 6485
+    },
+    {
+      "epoch": 17.306054557551562,
+      "grad_norm": 0.001404839102178812,
+      "learning_rate": 8.992909787404602e-06,
+      "loss": 0.017,
+      "step": 6490
+    },
+    {
+      "epoch": 17.31936127744511,
+      "grad_norm": 0.41923534870147705,
+      "learning_rate": 8.905433962469489e-06,
+      "loss": 0.0264,
+      "step": 6495
+    },
+    {
+      "epoch": 17.332667997338657,
+      "grad_norm": 0.00013469730038195848,
+      "learning_rate": 8.818365836066101e-06,
+      "loss": 0.0,
+      "step": 6500
+    },
+    {
+      "epoch": 17.345974717232203,
+      "grad_norm": 0.6122064590454102,
+      "learning_rate": 8.73170579787237e-06,
+      "loss": 0.0114,
+      "step": 6505
+    },
+    {
+      "epoch": 17.35928143712575,
+      "grad_norm": 1.719806432723999,
+      "learning_rate": 8.645454235739903e-06,
+      "loss": 0.045,
+      "step": 6510
+    },
+    {
+      "epoch": 17.372588157019294,
+      "grad_norm": 0.47664159536361694,
+      "learning_rate": 8.559611535692135e-06,
+      "loss": 0.0251,
+      "step": 6515
+    },
+    {
+      "epoch": 17.38589487691284,
+      "grad_norm": 9.255381883122027e-05,
+      "learning_rate": 8.474178081922524e-06,
+      "loss": 0.0424,
+      "step": 6520
+    },
+    {
+      "epoch": 17.39920159680639,
+      "grad_norm": 2.3919166778796352e-05,
+      "learning_rate": 8.38915425679304e-06,
+      "loss": 0.0126,
+      "step": 6525
+    },
+    {
+      "epoch": 17.412508316699935,
+      "grad_norm": 0.0001049041748046875,
+      "learning_rate": 8.304540440832298e-06,
+      "loss": 0.0189,
+      "step": 6530
+    },
+    {
+      "epoch": 17.42581503659348,
+      "grad_norm": 0.6109215617179871,
+      "learning_rate": 8.22033701273387e-06,
+      "loss": 0.0216,
+      "step": 6535
+    },
+    {
+      "epoch": 17.439121756487026,
+      "grad_norm": 9.519642480881885e-05,
+      "learning_rate": 8.13654434935467e-06,
+      "loss": 0.0244,
+      "step": 6540
+    },
+    {
+      "epoch": 17.45242847638057,
+      "grad_norm": 0.3934282660484314,
+      "learning_rate": 8.053162825713134e-06,
+      "loss": 0.0335,
+      "step": 6545
+    },
+    {
+      "epoch": 17.465735196274117,
+      "grad_norm": 0.0003310267929919064,
+      "learning_rate": 7.970192814987675e-06,
+      "loss": 0.0415,
+      "step": 6550
+    },
+    {
+      "epoch": 17.479041916167663,
+      "grad_norm": 6.156332528917119e-05,
+      "learning_rate": 7.887634688515e-06,
+      "loss": 0.0486,
+      "step": 6555
+    },
+    {
+      "epoch": 17.492348636061212,
+      "grad_norm": 0.0003013143432326615,
+      "learning_rate": 7.805488815788286e-06,
+      "loss": 0.0135,
+      "step": 6560
+    },
+    {
+      "epoch": 17.505655355954758,
+      "grad_norm": 5.3178453526925296e-05,
+      "learning_rate": 7.72375556445577e-06,
+      "loss": 0.0368,
+      "step": 6565
+    },
+    {
+      "epoch": 17.518962075848304,
+      "grad_norm": 3.9541144360555336e-05,
+      "learning_rate": 7.642435300318907e-06,
+      "loss": 0.0285,
+      "step": 6570
+    },
+    {
+      "epoch": 17.53226879574185,
+      "grad_norm": 0.5803095102310181,
+      "learning_rate": 7.561528387330796e-06,
+      "loss": 0.0201,
+      "step": 6575
+    },
+    {
+      "epoch": 17.545575515635395,
+      "grad_norm": 0.00011860304948640987,
+      "learning_rate": 7.48103518759462e-06,
+      "loss": 0.0253,
+      "step": 6580
+    },
+    {
+      "epoch": 17.55888223552894,
+      "grad_norm": 0.0002037498779827729,
+      "learning_rate": 7.400956061361974e-06,
+      "loss": 0.0279,
+      "step": 6585
+    },
+    {
+      "epoch": 17.57218895542249,
+      "grad_norm": 0.5323101878166199,
+      "learning_rate": 7.3212913670311355e-06,
+      "loss": 0.055,
+      "step": 6590
+    },
+    {
+      "epoch": 17.585495675316036,
+      "grad_norm": 0.00022871862165629864,
+      "learning_rate": 7.242041461145688e-06,
+      "loss": 0.0169,
+      "step": 6595
+    },
+    {
+      "epoch": 17.59880239520958,
+      "grad_norm": 0.0006196050089783967,
+      "learning_rate": 7.163206698392744e-06,
+      "loss": 0.0364,
+      "step": 6600
+    },
+    {
+      "epoch": 17.612109115103127,
+      "grad_norm": 0.7550682425498962,
+      "learning_rate": 7.084787431601436e-06,
+      "loss": 0.0423,
+      "step": 6605
+    },
+    {
+      "epoch": 17.625415834996673,
+      "grad_norm": 0.4589405655860901,
+      "learning_rate": 7.006784011741374e-06,
+      "loss": 0.0273,
+      "step": 6610
+    },
+    {
+      "epoch": 17.63872255489022,
+      "grad_norm": 0.7806951403617859,
+      "learning_rate": 6.929196787920899e-06,
+      "loss": 0.0095,
+      "step": 6615
+    },
+    {
+      "epoch": 17.652029274783764,
+      "grad_norm": 1.110194444656372,
+      "learning_rate": 6.852026107385756e-06,
+      "loss": 0.0538,
+      "step": 6620
+    },
+    {
+      "epoch": 17.665335994677314,
+      "grad_norm": 4.437299503479153e-05,
+      "learning_rate": 6.7752723155174226e-06,
+      "loss": 0.0419,
+      "step": 6625
+    },
+    {
+      "epoch": 17.67864271457086,
+      "grad_norm": 0.8215072154998779,
+      "learning_rate": 6.698935755831492e-06,
+      "loss": 0.0946,
+      "step": 6630
+    },
+    {
+      "epoch": 17.691949434464405,
+      "grad_norm": 0.46991631388664246,
+      "learning_rate": 6.623016769976309e-06,
+      "loss": 0.0473,
+      "step": 6635
+    },
+    {
+      "epoch": 17.70525615435795,
+      "grad_norm": 4.036861355416477e-05,
+      "learning_rate": 6.547515697731244e-06,
+      "loss": 0.0444,
+      "step": 6640
+    },
+    {
+      "epoch": 17.718562874251496,
+      "grad_norm": 0.4827462136745453,
+      "learning_rate": 6.472432877005341e-06,
+      "loss": 0.0585,
+      "step": 6645
+    },
+    {
+      "epoch": 17.731869594145042,
+      "grad_norm": 5.3903575462754816e-05,
+      "learning_rate": 6.397768643835755e-06,
+      "loss": 0.0099,
+      "step": 6650
+    },
+    {
+      "epoch": 17.745176314038588,
+      "grad_norm": 0.4768927991390228,
+      "learning_rate": 6.323523332386172e-06,
+      "loss": 0.0252,
+      "step": 6655
+    },
+    {
+      "epoch": 17.758483033932137,
+      "grad_norm": 1.0958563089370728,
+      "learning_rate": 6.2496972749453766e-06,
+      "loss": 0.0376,
+      "step": 6660
+    },
+    {
+      "epoch": 17.771789753825683,
+      "grad_norm": 0.0001639130205148831,
+      "learning_rate": 6.176290801925821e-06,
+      "loss": 0.0652,
+      "step": 6665
+    },
+    {
+      "epoch": 17.78509647371923,
+      "grad_norm": 0.7531500458717346,
+      "learning_rate": 6.103304241862007e-06,
+      "loss": 0.067,
+      "step": 6670
+    },
+    {
+      "epoch": 17.798403193612774,
+      "grad_norm": 0.6905338764190674,
+      "learning_rate": 6.030737921409169e-06,
+      "loss": 0.0175,
+      "step": 6675
+    },
+    {
+      "epoch": 17.81170991350632,
+      "grad_norm": 0.8327066898345947,
+      "learning_rate": 5.95859216534167e-06,
+      "loss": 0.0277,
+      "step": 6680
+    },
+    {
+      "epoch": 17.825016633399866,
+      "grad_norm": 0.8314510583877563,
+      "learning_rate": 5.886867296551646e-06,
+      "loss": 0.0669,
+      "step": 6685
+    },
+    {
+      "epoch": 17.83832335329341,
+      "grad_norm": 1.0047413110733032,
+      "learning_rate": 5.8155636360475385e-06,
+      "loss": 0.0529,
+      "step": 6690
+    },
+    {
+      "epoch": 17.85163007318696,
+      "grad_norm": 0.4971298575401306,
+      "learning_rate": 5.74468150295262e-06,
+      "loss": 0.0517,
+      "step": 6695
+    },
+    {
+      "epoch": 17.864936793080506,
+      "grad_norm": 6.307179864961654e-05,
+      "learning_rate": 5.674221214503639e-06,
+      "loss": 0.0346,
+      "step": 6700
+    },
+    {
+      "epoch": 17.878243512974052,
+      "grad_norm": 4.990035813534632e-05,
+      "learning_rate": 5.604183086049342e-06,
+      "loss": 0.0455,
+      "step": 6705
+    },
+    {
+      "epoch": 17.891550232867598,
+      "grad_norm": 0.7055619955062866,
+      "learning_rate": 5.534567431049009e-06,
+      "loss": 0.0293,
+      "step": 6710
+    },
+    {
+      "epoch": 17.904856952761143,
+      "grad_norm": 2.4250119167845696e-05,
+      "learning_rate": 5.465374561071212e-06,
+      "loss": 0.0179,
+      "step": 6715
+    },
+    {
+      "epoch": 17.91816367265469,
+      "grad_norm": 0.0003098284068983048,
+      "learning_rate": 5.396604785792281e-06,
+      "loss": 0.0296,
+      "step": 6720
+    },
+    {
+      "epoch": 17.93147039254824,
+      "grad_norm": 0.161666139960289,
+      "learning_rate": 5.328258412994958e-06,
+      "loss": 0.0405,
+      "step": 6725
+    },
+    {
+      "epoch": 17.944777112441784,
+      "grad_norm": 0.5785791873931885,
+      "learning_rate": 5.26033574856708e-06,
+      "loss": 0.0873,
+      "step": 6730
+    },
+    {
+      "epoch": 17.95808383233533,
+      "grad_norm": 0.1611766219139099,
+      "learning_rate": 5.192837096500058e-06,
+      "loss": 0.0396,
+      "step": 6735
+    },
+    {
+      "epoch": 17.971390552228875,
+      "grad_norm": 3.467617352725938e-05,
+      "learning_rate": 5.125762758887687e-06,
+      "loss": 0.0195,
+      "step": 6740
+    },
+    {
+      "epoch": 17.98469727212242,
+      "grad_norm": 0.00013225249131210148,
+      "learning_rate": 5.059113035924712e-06,
+      "loss": 0.0179,
+      "step": 6745
+    },
+    {
+      "epoch": 17.998003992015967,
+      "grad_norm": 0.4885573983192444,
+      "learning_rate": 4.992888225905468e-06,
+      "loss": 0.0431,
+      "step": 6750
+    },
+    {
+      "epoch": 17.998003992015967,
+      "eval_loss": 1.6460537910461426,
+      "eval_macro_f1": 68.10039104694555,
+      "eval_macro_precision": 70.49737286259203,
+      "eval_macro_recall": 66.37495253502834,
+      "eval_micro_f1": 87.5,
+      "eval_micro_precision": 87.5,
+      "eval_micro_recall": 87.5,
+      "eval_runtime": 4.834,
+      "eval_samples_per_second": 552.753,
+      "eval_steps_per_second": 34.547,
+      "step": 6750
+    },
+    {
+      "epoch": 18.013306719893546,
+      "grad_norm": 0.4641469419002533,
+      "learning_rate": 4.927088625222598e-06,
+      "loss": 0.0252,
+      "step": 6755
+    },
+    {
+      "epoch": 18.02661343978709,
+      "grad_norm": 0.00014642822497989982,
+      "learning_rate": 4.861714528365646e-06,
+      "loss": 0.0091,
+      "step": 6760
+    },
+    {
+      "epoch": 18.039920159680637,
+      "grad_norm": 3.572425339370966e-05,
+      "learning_rate": 4.796766227919857e-06,
+      "loss": 0.0282,
+      "step": 6765
+    },
+    {
+      "epoch": 18.053226879574186,
+      "grad_norm": 0.449663907289505,
+      "learning_rate": 4.7322440145647905e-06,
+      "loss": 0.0163,
+      "step": 6770
+    },
+    {
+      "epoch": 18.066533599467732,
+      "grad_norm": 0.6586659550666809,
+      "learning_rate": 4.668148177072984e-06,
+      "loss": 0.0358,
+      "step": 6775
+    },
+    {
+      "epoch": 18.079840319361278,
+      "grad_norm": 0.6322921514511108,
+      "learning_rate": 4.604479002308737e-06,
+      "loss": 0.0335,
+      "step": 6780
+    },
+    {
+      "epoch": 18.093147039254823,
+      "grad_norm": 7.740201544947922e-05,
+      "learning_rate": 4.541236775226809e-06,
+      "loss": 0.0344,
+      "step": 6785
+    },
+    {
+      "epoch": 18.10645375914837,
+      "grad_norm": 7.509323040721938e-05,
+      "learning_rate": 4.478421778871112e-06,
+      "loss": 0.0001,
+      "step": 6790
+    },
+    {
+      "epoch": 18.119760479041915,
+      "grad_norm": 0.0002992966619785875,
+      "learning_rate": 4.416034294373472e-06,
+      "loss": 0.0082,
+      "step": 6795
+    },
+    {
+      "epoch": 18.133067198935464,
+      "grad_norm": 0.2723856270313263,
+      "learning_rate": 4.354074600952407e-06,
+      "loss": 0.0689,
+      "step": 6800
+    },
+    {
+      "epoch": 18.14637391882901,
+      "grad_norm": 0.7682840824127197,
+      "learning_rate": 4.292542975911717e-06,
+      "loss": 0.0656,
+      "step": 6805
+    },
+    {
+      "epoch": 18.159680638722556,
+      "grad_norm": 0.5859367847442627,
+      "learning_rate": 4.231439694639483e-06,
+      "loss": 0.0403,
+      "step": 6810
+    },
+    {
+      "epoch": 18.1729873586161,
+      "grad_norm": 0.0007575666531920433,
+      "learning_rate": 4.170765030606638e-06,
+      "loss": 0.0266,
+      "step": 6815
+    },
+    {
+      "epoch": 18.186294078509647,
+      "grad_norm": 0.83262038230896,
+      "learning_rate": 4.110519255365852e-06,
+      "loss": 0.0619,
+      "step": 6820
+    },
+    {
+      "epoch": 18.199600798403193,
+      "grad_norm": 0.00012123231863370165,
+      "learning_rate": 4.050702638550275e-06,
+      "loss": 0.0531,
+      "step": 6825
+    },
+    {
+      "epoch": 18.21290751829674,
+      "grad_norm": 0.47834277153015137,
+      "learning_rate": 3.991315447872302e-06,
+      "loss": 0.029,
+      "step": 6830
+    },
+    {
+      "epoch": 18.226214238190288,
+      "grad_norm": 0.0005108633195050061,
+      "learning_rate": 3.932357949122445e-06,
+      "loss": 0.0282,
+      "step": 6835
+    },
+    {
+      "epoch": 18.239520958083833,
+      "grad_norm": 8.286790398415178e-05,
+      "learning_rate": 3.873830406168111e-06,
+      "loss": 0.0206,
+      "step": 6840
+    },
+    {
+      "epoch": 18.25282767797738,
+      "grad_norm": 0.4972341060638428,
+      "learning_rate": 3.815733080952411e-06,
+      "loss": 0.0435,
+      "step": 6845
+    },
+    {
+      "epoch": 18.266134397870925,
+      "grad_norm": 0.00017122157441917807,
+      "learning_rate": 3.758066233492952e-06,
+      "loss": 0.0351,
+      "step": 6850
+    },
+    {
+      "epoch": 18.27944111776447,
+      "grad_norm": 0.5052186250686646,
+      "learning_rate": 3.7008301218807716e-06,
+      "loss": 0.0492,
+      "step": 6855
+    },
+    {
+      "epoch": 18.292747837658016,
+      "grad_norm": 0.3930492401123047,
+      "learning_rate": 3.6440250022790834e-06,
+      "loss": 0.0149,
+      "step": 6860
+    },
+    {
+      "epoch": 18.306054557551562,
+      "grad_norm": 0.00020653626415878534,
+      "learning_rate": 3.5876511289222314e-06,
+      "loss": 0.029,
+      "step": 6865
+    },
+    {
+      "epoch": 18.31936127744511,
+      "grad_norm": 0.001036687521263957,
+      "learning_rate": 3.5317087541144377e-06,
+      "loss": 0.0627,
+      "step": 6870
+    },
+    {
+      "epoch": 18.332667997338657,
+      "grad_norm": 0.1497216373682022,
+      "learning_rate": 3.476198128228736e-06,
+      "loss": 0.0134,
+      "step": 6875
+    },
+    {
+      "epoch": 18.345974717232203,
+      "grad_norm": 0.003397771855816245,
+      "learning_rate": 3.4211194997058714e-06,
+      "loss": 0.0346,
+      "step": 6880
+    },
+    {
+      "epoch": 18.35928143712575,
+      "grad_norm": 0.001260089105926454,
+      "learning_rate": 3.3664731150531482e-06,
+      "loss": 0.0244,
+      "step": 6885
+    },
+    {
+      "epoch": 18.372588157019294,
+      "grad_norm": 0.49642452597618103,
+      "learning_rate": 3.3122592188433497e-06,
+      "loss": 0.0461,
+      "step": 6890
+    },
+    {
+      "epoch": 18.38589487691284,
+      "grad_norm": 0.519691526889801,
+      "learning_rate": 3.2584780537136207e-06,
+      "loss": 0.0808,
+      "step": 6895
+    },
+    {
+      "epoch": 18.39920159680639,
+      "grad_norm": 0.0005471158656291664,
+      "learning_rate": 3.2051298603643753e-06,
+      "loss": 0.0,
+      "step": 6900
+    },
+    {
+      "epoch": 18.412508316699935,
+      "grad_norm": 0.5794316530227661,
+      "learning_rate": 3.1522148775583016e-06,
+      "loss": 0.0759,
+      "step": 6905
+    },
+    {
+      "epoch": 18.42581503659348,
+      "grad_norm": 0.00023820818751119077,
+      "learning_rate": 3.099733342119171e-06,
+      "loss": 0.0187,
+      "step": 6910
+    },
+    {
+      "epoch": 18.439121756487026,
+      "grad_norm": 0.7367857694625854,
+      "learning_rate": 3.047685488930874e-06,
+      "loss": 0.0367,
+      "step": 6915
+    },
+    {
+      "epoch": 18.45242847638057,
+      "grad_norm": 0.7412865161895752,
+      "learning_rate": 2.996071550936319e-06,
+      "loss": 0.0626,
+      "step": 6920
+    },
+    {
+      "epoch": 18.465735196274117,
+      "grad_norm": 0.49905386567115784,
+      "learning_rate": 2.9448917591363925e-06,
+      "loss": 0.0273,
+      "step": 6925
+    },
+    {
+      "epoch": 18.479041916167663,
+      "grad_norm": 0.7662301063537598,
+      "learning_rate": 2.894146342588977e-06,
+      "loss": 0.0338,
+      "step": 6930
+    },
+    {
+      "epoch": 18.492348636061212,
+      "grad_norm": 7.189960888354108e-05,
+      "learning_rate": 2.8438355284078234e-06,
+      "loss": 0.0461,
+      "step": 6935
+    },
+    {
+      "epoch": 18.505655355954758,
+      "grad_norm": 0.00035782987833954394,
+      "learning_rate": 2.793959541761659e-06,
+      "loss": 0.0098,
+      "step": 6940
+    },
+    {
+      "epoch": 18.518962075848304,
+      "grad_norm": 0.0002716764574870467,
+      "learning_rate": 2.744518605873092e-06,
+      "loss": 0.0429,
+      "step": 6945
+    },
+    {
+      "epoch": 18.53226879574185,
+      "grad_norm": 0.5008606314659119,
+      "learning_rate": 2.6955129420176196e-06,
+      "loss": 0.0341,
+      "step": 6950
+    },
+    {
+      "epoch": 18.545575515635395,
+      "grad_norm": 0.6177487969398499,
+      "learning_rate": 2.6469427695226757e-06,
+      "loss": 0.0268,
+      "step": 6955
+    },
+    {
+      "epoch": 18.55888223552894,
+      "grad_norm": 9.384102304466069e-05,
+      "learning_rate": 2.5988083057666533e-06,
+      "loss": 0.0555,
+      "step": 6960
+    },
+    {
+      "epoch": 18.57218895542249,
+      "grad_norm": 0.00016950270219240338,
+      "learning_rate": 2.5511097661778503e-06,
+      "loss": 0.0441,
+      "step": 6965
+    },
+    {
+      "epoch": 18.585495675316036,
+      "grad_norm": 0.5722589492797852,
+      "learning_rate": 2.5038473642336137e-06,
+      "loss": 0.0268,
+      "step": 6970
+    },
+    {
+      "epoch": 18.59880239520958,
+      "grad_norm": 0.5383985042572021,
+      "learning_rate": 2.4570213114592954e-06,
+      "loss": 0.0515,
+      "step": 6975
+    },
+    {
+      "epoch": 18.612109115103127,
+      "grad_norm": 0.00039173016557469964,
+      "learning_rate": 2.410631817427378e-06,
+      "loss": 0.0372,
+      "step": 6980
+    },
+    {
+      "epoch": 18.625415834996673,
+      "grad_norm": 0.6350005269050598,
+      "learning_rate": 2.3646790897564943e-06,
+      "loss": 0.0265,
+      "step": 6985
+    },
+    {
+      "epoch": 18.63872255489022,
+      "grad_norm": 0.00014799917698837817,
+      "learning_rate": 2.3191633341104856e-06,
+      "loss": 0.0177,
+      "step": 6990
+    },
+    {
+      "epoch": 18.652029274783764,
+      "grad_norm": 0.0001857498282333836,
+      "learning_rate": 2.274084754197514e-06,
+      "loss": 0.0408,
+      "step": 6995
+    },
+    {
+      "epoch": 18.665335994677314,
+      "grad_norm": 0.00058632658328861,
+      "learning_rate": 2.2294435517691503e-06,
+      "loss": 0.0175,
+      "step": 7000
+    },
+    {
+      "epoch": 18.67864271457086,
+      "grad_norm": 0.0031339959241449833,
+      "learning_rate": 2.1852399266194314e-06,
+      "loss": 0.0115,
+      "step": 7005
+    },
+    {
+      "epoch": 18.691949434464405,
+      "grad_norm": 8.091120253084227e-05,
+      "learning_rate": 2.141474076584038e-06,
+      "loss": 0.0221,
+      "step": 7010
+    },
+    {
+      "epoch": 18.70525615435795,
+      "grad_norm": 0.013986263424158096,
+      "learning_rate": 2.098146197539319e-06,
+      "loss": 0.0082,
+      "step": 7015
+    },
+    {
+      "epoch": 18.718562874251496,
+      "grad_norm": 0.5572426319122314,
+      "learning_rate": 2.05525648340148e-06,
+      "loss": 0.0353,
+      "step": 7020
+    },
+    {
+      "epoch": 18.731869594145042,
+      "grad_norm": 7.861760241212323e-05,
+      "learning_rate": 2.0128051261257164e-06,
+      "loss": 0.0,
+      "step": 7025
+    },
+    {
+      "epoch": 18.745176314038588,
+      "grad_norm": 6.267589196795598e-05,
+      "learning_rate": 1.9707923157052833e-06,
+      "loss": 0.0348,
+      "step": 7030
+    },
+    {
+      "epoch": 18.758483033932137,
+      "grad_norm": 0.5337706804275513,
+      "learning_rate": 1.9292182401707603e-06,
+      "loss": 0.0535,
+      "step": 7035
+    },
+    {
+      "epoch": 18.771789753825683,
+      "grad_norm": 0.5486602783203125,
+      "learning_rate": 1.8880830855891096e-06,
+      "loss": 0.0181,
+      "step": 7040
+    },
+    {
+      "epoch": 18.78509647371923,
+      "grad_norm": 0.42544683814048767,
+      "learning_rate": 1.847387036062853e-06,
+      "loss": 0.0521,
+      "step": 7045
+    },
+    {
+      "epoch": 18.798403193612774,
+      "grad_norm": 0.553593099117279,
+      "learning_rate": 1.8071302737293295e-06,
+      "loss": 0.0457,
+      "step": 7050
+    },
+    {
+      "epoch": 18.81170991350632,
+      "grad_norm": 0.00020910767489112914,
+      "learning_rate": 1.7673129787598054e-06,
+      "loss": 0.0242,
+      "step": 7055
+    },
+    {
+      "epoch": 18.825016633399866,
+      "grad_norm": 0.00025782870943658054,
+      "learning_rate": 1.7279353293586765e-06,
+      "loss": 0.0356,
+      "step": 7060
+    },
+    {
+      "epoch": 18.83832335329341,
+      "grad_norm": 7.453611033270136e-05,
+      "learning_rate": 1.6889975017626903e-06,
+      "loss": 0.035,
+      "step": 7065
+    },
+    {
+      "epoch": 18.85163007318696,
+      "grad_norm": 0.6055145859718323,
+      "learning_rate": 1.6504996702401243e-06,
+      "loss": 0.0375,
+      "step": 7070
+    },
+    {
+      "epoch": 18.864936793080506,
+      "grad_norm": 0.5281118750572205,
+      "learning_rate": 1.6124420070900758e-06,
+      "loss": 0.0345,
+      "step": 7075
+    },
+    {
+      "epoch": 18.878243512974052,
+      "grad_norm": 0.5610820055007935,
+      "learning_rate": 1.574824682641629e-06,
+      "loss": 0.0539,
+      "step": 7080
+    },
+    {
+      "epoch": 18.891550232867598,
+      "grad_norm": 1.1242096424102783,
+      "learning_rate": 1.537647865253089e-06,
+      "loss": 0.0472,
+      "step": 7085
+    },
+    {
+      "epoch": 18.904856952761143,
+      "grad_norm": 1.0025428533554077,
+      "learning_rate": 1.50091172131126e-06,
+      "loss": 0.0549,
+      "step": 7090
+    },
+    {
+      "epoch": 18.91816367265469,
+      "grad_norm": 0.46737486124038696,
+      "learning_rate": 1.4646164152307018e-06,
+      "loss": 0.0545,
+      "step": 7095
+    },
+    {
+      "epoch": 18.93147039254824,
+      "grad_norm": 0.5139942765235901,
+      "learning_rate": 1.4287621094529524e-06,
+      "loss": 0.0264,
+      "step": 7100
+    },
+    {
+      "epoch": 18.944777112441784,
+      "grad_norm": 3.634204404079355e-05,
+      "learning_rate": 1.3933489644458619e-06,
+      "loss": 0.0,
+      "step": 7105
+    },
+    {
+      "epoch": 18.95808383233533,
+      "grad_norm": 1.1411399841308594,
+      "learning_rate": 1.3583771387028265e-06,
+      "loss": 0.0495,
+      "step": 7110
+    },
+    {
+      "epoch": 18.971390552228875,
+      "grad_norm": 0.0005110076745040715,
+      "learning_rate": 1.323846788742078e-06,
+      "loss": 0.0249,
+      "step": 7115
+    },
+    {
+      "epoch": 18.98469727212242,
+      "grad_norm": 0.8430808186531067,
+      "learning_rate": 1.2897580691060506e-06,
+      "loss": 0.0498,
+      "step": 7120
+    },
+    {
+      "epoch": 18.998003992015967,
+      "grad_norm": 9.565344225848094e-05,
+      "learning_rate": 1.2561111323605712e-06,
+      "loss": 0.0094,
+      "step": 7125
+    },
+    {
+      "epoch": 18.998003992015967,
+      "eval_loss": 1.6504560708999634,
+      "eval_macro_f1": 68.05326783992479,
+      "eval_macro_precision": 70.41695830446321,
+      "eval_macro_recall": 66.35973183487613,
+      "eval_micro_f1": 87.4625748502994,
+      "eval_micro_precision": 87.4625748502994,
+      "eval_micro_recall": 87.4625748502994,
+      "eval_runtime": 4.4506,
+      "eval_samples_per_second": 600.367,
+      "eval_steps_per_second": 37.523,
+      "step": 7125
+    },
+    {
+      "epoch": 19.013306719893546,
+      "grad_norm": 4.465243182494305e-05,
+      "learning_rate": 1.2229061290942922e-06,
+      "loss": 0.0366,
+      "step": 7130
+    },
+    {
+      "epoch": 19.02661343978709,
+      "grad_norm": 0.0008600103319622576,
+      "learning_rate": 1.1901432079179709e-06,
+      "loss": 0.0187,
+      "step": 7135
+    },
+    {
+      "epoch": 19.039920159680637,
+      "grad_norm": 4.5936063543194905e-05,
+      "learning_rate": 1.157822515463758e-06,
+      "loss": 0.0644,
+      "step": 7140
+    },
+    {
+      "epoch": 19.053226879574186,
+      "grad_norm": 0.1492280811071396,
+      "learning_rate": 1.1259441963846429e-06,
+      "loss": 0.0575,
+      "step": 7145
+    },
+    {
+      "epoch": 19.066533599467732,
+      "grad_norm": 8.396996418014169e-05,
+      "learning_rate": 1.0945083933537103e-06,
+      "loss": 0.0452,
+      "step": 7150
+    },
+    {
+      "epoch": 19.079840319361278,
+      "grad_norm": 0.4817587435245514,
+      "learning_rate": 1.0635152470635512e-06,
+      "loss": 0.0221,
+      "step": 7155
+    },
+    {
+      "epoch": 19.093147039254823,
+      "grad_norm": 8.231854008045048e-05,
+      "learning_rate": 1.0329648962256411e-06,
+      "loss": 0.0077,
+      "step": 7160
+    },
+    {
+      "epoch": 19.10645375914837,
+      "grad_norm": 5.4089796321932226e-05,
+      "learning_rate": 1.002857477569663e-06,
+      "loss": 0.0174,
+      "step": 7165
+    },
+    {
+      "epoch": 19.119760479041915,
+      "grad_norm": 0.20306649804115295,
+      "learning_rate": 9.731931258429638e-07,
+      "loss": 0.024,
+      "step": 7170
+    },
+    {
+      "epoch": 19.133067198935464,
+      "grad_norm": 0.4647962152957916,
+      "learning_rate": 9.439719738099317e-07,
+      "loss": 0.0156,
+      "step": 7175
+    },
+    {
+      "epoch": 19.14637391882901,
+      "grad_norm": 5.901711483602412e-05,
+      "learning_rate": 9.15194152251353e-07,
+      "loss": 0.0333,
+      "step": 7180
+    },
+    {
+      "epoch": 19.159680638722556,
+      "grad_norm": 0.00029276739223860204,
+      "learning_rate": 8.868597899638898e-07,
+      "loss": 0.0171,
+      "step": 7185
+    },
+    {
+      "epoch": 19.1729873586161,
+      "grad_norm": 0.46516790986061096,
+      "learning_rate": 8.589690137594697e-07,
+      "loss": 0.0254,
+      "step": 7190
+    },
+    {
+      "epoch": 19.186294078509647,
+      "grad_norm": 0.00011026985885109752,
+      "learning_rate": 8.315219484647308e-07,
+      "loss": 0.0425,
+      "step": 7195
+    },
+    {
+      "epoch": 19.199600798403193,
+      "grad_norm": 5.936959132668562e-05,
+      "learning_rate": 8.04518716920466e-07,
+      "loss": 0.0362,
+      "step": 7200
+    },
+    {
+      "epoch": 19.21290751829674,
+      "grad_norm": 3.622892836574465e-05,
+      "learning_rate": 7.779594399810685e-07,
+      "loss": 0.0082,
+      "step": 7205
+    },
+    {
+      "epoch": 19.226214238190288,
+      "grad_norm": 0.00019010104006156325,
+      "learning_rate": 7.518442365139766e-07,
+      "loss": 0.0529,
+      "step": 7210
+    },
+    {
+      "epoch": 19.239520958083833,
+      "grad_norm": 6.970770482439548e-05,
+      "learning_rate": 7.261732233991513e-07,
+      "loss": 0.0409,
+      "step": 7215
+    },
+    {
+      "epoch": 19.25282767797738,
+      "grad_norm": 5.463602428790182e-05,
+      "learning_rate": 7.009465155285777e-07,
+      "loss": 0.048,
+      "step": 7220
+    },
+    {
+      "epoch": 19.266134397870925,
+      "grad_norm": 0.7579261660575867,
+      "learning_rate": 6.761642258056978e-07,
+      "loss": 0.0654,
+      "step": 7225
+    },
+    {
+      "epoch": 19.27944111776447,
+      "grad_norm": 0.48954853415489197,
+      "learning_rate": 6.518264651449779e-07,
+      "loss": 0.0352,
+      "step": 7230
+    },
+    {
+      "epoch": 19.292747837658016,
+      "grad_norm": 3.727992225321941e-05,
+      "learning_rate": 6.279333424713429e-07,
+      "loss": 0.0436,
+      "step": 7235
+    },
+    {
+      "epoch": 19.306054557551562,
+      "grad_norm": 4.1896390030160546e-05,
+      "learning_rate": 6.044849647197093e-07,
+      "loss": 0.0293,
+      "step": 7240
+    },
+    {
+      "epoch": 19.31936127744511,
+      "grad_norm": 0.5448512434959412,
+      "learning_rate": 5.814814368345412e-07,
+      "loss": 0.0343,
+      "step": 7245
+    },
+    {
+      "epoch": 19.332667997338657,
+      "grad_norm": 0.5785821676254272,
+      "learning_rate": 5.589228617693288e-07,
+      "loss": 0.0362,
+      "step": 7250
+    },
+    {
+      "epoch": 19.345974717232203,
+      "grad_norm": 0.547781765460968,
+      "learning_rate": 5.368093404861774e-07,
+      "loss": 0.0145,
+      "step": 7255
+    },
+    {
+      "epoch": 19.35928143712575,
+      "grad_norm": 3.6774767067981884e-05,
+      "learning_rate": 5.151409719553079e-07,
+      "loss": 0.0167,
+      "step": 7260
+    },
+    {
+      "epoch": 19.372588157019294,
+      "grad_norm": 5.5066615459509194e-05,
+      "learning_rate": 4.939178531546462e-07,
+      "loss": 0.0351,
+      "step": 7265
+    },
+    {
+      "epoch": 19.38589487691284,
+      "grad_norm": 4.909703784505837e-05,
+      "learning_rate": 4.731400790693785e-07,
+      "loss": 0.0171,
+      "step": 7270
+    },
+    {
+      "epoch": 19.39920159680639,
+      "grad_norm": 0.5192339420318604,
+      "learning_rate": 4.5280774269154115e-07,
+      "loss": 0.0357,
+      "step": 7275
+    },
+    {
+      "epoch": 19.412508316699935,
+      "grad_norm": 0.5900948643684387,
+      "learning_rate": 4.329209350195651e-07,
+      "loss": 0.0571,
+      "step": 7280
+    },
+    {
+      "epoch": 19.42581503659348,
+      "grad_norm": 1.5432310104370117,
+      "learning_rate": 4.1347974505789867e-07,
+      "loss": 0.0472,
+      "step": 7285
+    },
+    {
+      "epoch": 19.439121756487026,
+      "grad_norm": 0.0005136824329383671,
+      "learning_rate": 3.9448425981661876e-07,
+      "loss": 0.0274,
+      "step": 7290
+    },
+    {
+      "epoch": 19.45242847638057,
+      "grad_norm": 0.6122019290924072,
+      "learning_rate": 3.7593456431103123e-07,
+      "loss": 0.0292,
+      "step": 7295
+    },
+    {
+      "epoch": 19.465735196274117,
+      "grad_norm": 0.0002864995039999485,
+      "learning_rate": 3.578307415612714e-07,
+      "loss": 0.0358,
+      "step": 7300
+    },
+    {
+      "epoch": 19.479041916167663,
+      "grad_norm": 1.156638503074646,
+      "learning_rate": 3.401728725919373e-07,
+      "loss": 0.047,
+      "step": 7305
+    },
+    {
+      "epoch": 19.492348636061212,
+      "grad_norm": 0.7358021140098572,
+      "learning_rate": 3.229610364317792e-07,
+      "loss": 0.0536,
+      "step": 7310
+    },
+    {
+      "epoch": 19.505655355954758,
+      "grad_norm": 0.0011500949040055275,
+      "learning_rate": 3.061953101132442e-07,
+      "loss": 0.0213,
+      "step": 7315
+    },
+    {
+      "epoch": 19.518962075848304,
+      "grad_norm": 0.4783131182193756,
+      "learning_rate": 2.898757686722542e-07,
+      "loss": 0.0264,
+      "step": 7320
+    },
+    {
+      "epoch": 19.53226879574185,
+      "grad_norm": 0.00014955556252971292,
+      "learning_rate": 2.7400248514776183e-07,
+      "loss": 0.0446,
+      "step": 7325
+    },
+    {
+      "epoch": 19.545575515635395,
+      "grad_norm": 0.0015204122755676508,
+      "learning_rate": 2.585755305814841e-07,
+      "loss": 0.0029,
+      "step": 7330
+    },
+    {
+      "epoch": 19.55888223552894,
+      "grad_norm": 3.2041556551121175e-05,
+      "learning_rate": 2.4359497401758024e-07,
+      "loss": 0.0165,
+      "step": 7335
+    },
+    {
+      "epoch": 19.57218895542249,
+      "grad_norm": 4.298180647310801e-05,
+      "learning_rate": 2.2906088250229664e-07,
+      "loss": 0.028,
+      "step": 7340
+    },
+    {
+      "epoch": 19.585495675316036,
+      "grad_norm": 4.960396108799614e-05,
+      "learning_rate": 2.1497332108375568e-07,
+      "loss": 0.0264,
+      "step": 7345
+    },
+    {
+      "epoch": 19.59880239520958,
+      "grad_norm": 7.29636667529121e-05,
+      "learning_rate": 2.0133235281156736e-07,
+      "loss": 0.0161,
+      "step": 7350
+    },
+    {
+      "epoch": 19.612109115103127,
+      "grad_norm": 0.5350423455238342,
+      "learning_rate": 1.8813803873659607e-07,
+      "loss": 0.0601,
+      "step": 7355
+    },
+    {
+      "epoch": 19.625415834996673,
+      "grad_norm": 2.966666579595767e-05,
+      "learning_rate": 1.753904379106941e-07,
+      "loss": 0.0095,
+      "step": 7360
+    },
+    {
+      "epoch": 19.63872255489022,
+      "grad_norm": 0.0019933367148041725,
+      "learning_rate": 1.630896073864352e-07,
+      "loss": 0.0165,
+      "step": 7365
+    },
+    {
+      "epoch": 19.652029274783764,
+      "grad_norm": 0.00011221903696423396,
+      "learning_rate": 1.5123560221681487e-07,
+      "loss": 0.0029,
+      "step": 7370
+    },
+    {
+      "epoch": 19.665335994677314,
+      "grad_norm": 0.0003375353990122676,
+      "learning_rate": 1.3982847545507271e-07,
+      "loss": 0.0397,
+      "step": 7375
+    },
+    {
+      "epoch": 19.67864271457086,
+      "grad_norm": 0.9430463910102844,
+      "learning_rate": 1.2886827815440372e-07,
+      "loss": 0.0417,
+      "step": 7380
+    },
+    {
+      "epoch": 19.691949434464405,
+      "grad_norm": 0.1029142215847969,
+      "learning_rate": 1.1835505936773628e-07,
+      "loss": 0.0372,
+      "step": 7385
+    },
+    {
+      "epoch": 19.70525615435795,
+      "grad_norm": 0.0002183911856263876,
+      "learning_rate": 1.0828886614754341e-07,
+      "loss": 0.0246,
+      "step": 7390
+    },
+    {
+      "epoch": 19.718562874251496,
+      "grad_norm": 0.5175012946128845,
+      "learning_rate": 9.866974354560965e-08,
+      "loss": 0.0408,
+      "step": 7395
+    },
+    {
+      "epoch": 19.731869594145042,
+      "grad_norm": 0.14726030826568604,
+      "learning_rate": 8.949773461282008e-08,
+      "loss": 0.0144,
+      "step": 7400
+    },
+    {
+      "epoch": 19.745176314038588,
+      "grad_norm": 6.141362973721698e-05,
+      "learning_rate": 8.077288039898267e-08,
+      "loss": 0.0249,
+      "step": 7405
+    },
+    {
+      "epoch": 19.758483033932137,
+      "grad_norm": 0.4466927647590637,
+      "learning_rate": 7.249521995263964e-08,
+      "loss": 0.022,
+      "step": 7410
+    },
+    {
+      "epoch": 19.771789753825683,
+      "grad_norm": 0.4848124384880066,
+      "learning_rate": 6.466479032091189e-08,
+      "loss": 0.0246,
+      "step": 7415
+    },
+    {
+      "epoch": 19.78509647371923,
+      "grad_norm": 0.5077440738677979,
+      "learning_rate": 5.7281626549277046e-08,
+      "loss": 0.0317,
+      "step": 7420
+    },
+    {
+      "epoch": 19.798403193612774,
+      "grad_norm": 0.0002426803985144943,
+      "learning_rate": 5.0345761681491746e-08,
+      "loss": 0.0332,
+      "step": 7425
+    },
+    {
+      "epoch": 19.81170991350632,
+      "grad_norm": 0.7584611177444458,
+      "learning_rate": 4.385722675936954e-08,
+      "loss": 0.0657,
+      "step": 7430
+    },
+    {
+      "epoch": 19.825016633399866,
+      "grad_norm": 0.4777810573577881,
+      "learning_rate": 3.781605082270323e-08,
+      "loss": 0.0693,
+      "step": 7435
+    },
+    {
+      "epoch": 19.83832335329341,
+      "grad_norm": 0.4914855659008026,
+      "learning_rate": 3.2222260909087196e-08,
+      "loss": 0.0568,
+      "step": 7440
+    },
+    {
+      "epoch": 19.85163007318696,
+      "grad_norm": 0.14886783063411713,
+      "learning_rate": 2.7075882053828605e-08,
+      "loss": 0.0375,
+      "step": 7445
+    },
+    {
+      "epoch": 19.864936793080506,
+      "grad_norm": 0.5256046056747437,
+      "learning_rate": 2.2376937289814158e-08,
+      "loss": 0.0281,
+      "step": 7450
+    },
+    {
+      "epoch": 19.878243512974052,
+      "grad_norm": 0.5002803802490234,
+      "learning_rate": 1.81254476474213e-08,
+      "loss": 0.0512,
+      "step": 7455
+    },
+    {
+      "epoch": 19.891550232867598,
+      "grad_norm": 0.5000232458114624,
+      "learning_rate": 1.4321432154418279e-08,
+      "loss": 0.0741,
+      "step": 7460
+    },
+    {
+      "epoch": 19.904856952761143,
+      "grad_norm": 0.00015142855409067124,
+      "learning_rate": 1.0964907835864235e-08,
+      "loss": 0.0367,
+      "step": 7465
+    },
+    {
+      "epoch": 19.91816367265469,
+      "grad_norm": 0.7019768357276917,
+      "learning_rate": 8.055889714064791e-09,
+      "loss": 0.0499,
+      "step": 7470
+    },
+    {
+      "epoch": 19.93147039254824,
+      "grad_norm": 4.4128122681286186e-05,
+      "learning_rate": 5.594390808494332e-09,
+      "loss": 0.035,
+      "step": 7475
+    },
+    {
+      "epoch": 19.944777112441784,
+      "grad_norm": 0.5177294015884399,
+      "learning_rate": 3.5804221357182976e-09,
+      "loss": 0.0339,
+      "step": 7480
+    },
+    {
+      "epoch": 19.95808383233533,
+      "grad_norm": 0.0001399219618178904,
+      "learning_rate": 2.0139927093487664e-09,
+      "loss": 0.0168,
+      "step": 7485
+    },
+    {
+      "epoch": 19.971390552228875,
+      "grad_norm": 0.8865404725074768,
+      "learning_rate": 8.951095400333564e-10,
+      "loss": 0.0492,
+      "step": 7490
+    },
+    {
+      "epoch": 19.98469727212242,
+      "grad_norm": 0.43994656205177307,
+      "learning_rate": 2.237776353886112e-10,
+      "loss": 0.0501,
+      "step": 7495
+    },
+    {
+      "epoch": 19.998003992015967,
+      "grad_norm": 0.00010475648014107719,
+      "learning_rate": 0.0,
+      "loss": 0.0174,
+      "step": 7500
+    },
+    {
+      "epoch": 19.998003992015967,
+      "eval_loss": 1.651909589767456,
+      "eval_macro_f1": 68.1809240430323,
+      "eval_macro_precision": 70.57937372214415,
+      "eval_macro_recall": 66.4831886249996,
+      "eval_micro_f1": 87.5,
+      "eval_micro_precision": 87.5,
+      "eval_micro_recall": 87.5,
+      "eval_runtime": 5.4153,
+      "eval_samples_per_second": 493.416,
+      "eval_steps_per_second": 30.838,
+      "step": 7500
+    },
+    {
+      "epoch": 19.998003992015967,
+      "step": 7500,
+      "total_flos": 1.5762278746161152e+16,
+      "train_loss": 0.23192051490644613,
+      "train_runtime": 3495.581,
+      "train_samples_per_second": 137.579,
+      "train_steps_per_second": 2.146
     }
   ],
   "logging_steps": 5,
-  "max_steps": 1580,
+  "max_steps": 7500,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 20,
   "save_steps": 50.0,
@@ -2528,7 +10816,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 8695327329615872.0,
+  "total_flos": 1.5762278746161152e+16,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null