diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -10,9 +10,9 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 1.3235913515090942,
+      "grad_norm": 1.4515146017074585,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 1.0641,
+      "loss": 1.0545,
       "step": 96
     },
     {
@@ -29,21 +29,21 @@
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.0,
       "eval_PERSON_recall": 0.0,
-      "eval_loss": 0.6655052900314331,
+      "eval_loss": 0.6621571183204651,
       "eval_overall_accuracy": 0.8394180620367828,
       "eval_overall_f1": 0.0,
       "eval_overall_precision": 0.0,
       "eval_overall_recall": 0.0,
-      "eval_runtime": 0.614,
-      "eval_samples_per_second": 276.861,
-      "eval_steps_per_second": 4.886,
+      "eval_runtime": 0.2785,
+      "eval_samples_per_second": 610.49,
+      "eval_steps_per_second": 10.773,
       "step": 96
     },
     {
       "epoch": 2.0,
-      "grad_norm": 1.5239665508270264,
+      "grad_norm": 2.103937864303589,
       "learning_rate": 4.9e-05,
-      "loss": 0.6402,
+      "loss": 0.64,
       "step": 192
     },
     {
@@ -60,835 +60,835 @@
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.0,
       "eval_PERSON_recall": 0.0,
-      "eval_loss": 0.5235269665718079,
+      "eval_loss": 0.5205754637718201,
       "eval_overall_accuracy": 0.8399670601152895,
       "eval_overall_f1": 0.009900990099009901,
       "eval_overall_precision": 0.3333333333333333,
       "eval_overall_recall": 0.005025125628140704,
-      "eval_runtime": 0.6248,
-      "eval_samples_per_second": 272.086,
-      "eval_steps_per_second": 4.802,
+      "eval_runtime": 0.2762,
+      "eval_samples_per_second": 615.56,
+      "eval_steps_per_second": 10.863,
       "step": 192
     },
     {
       "epoch": 3.0,
-      "grad_norm": 0.9175276160240173,
+      "grad_norm": 0.8918240666389465,
       "learning_rate": 4.85e-05,
-      "loss": 0.5003,
+      "loss": 0.503,
       "step": 288
     },
     {
       "epoch": 3.0,
-      "eval_LOCATION_f1": 0.03076923076923077,
+      "eval_LOCATION_f1": 0.015625,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.14285714285714285,
-      "eval_LOCATION_recall": 0.017241379310344827,
-      "eval_ORGANIZATION_f1": 0.25,
+      "eval_LOCATION_precision": 0.08333333333333333,
+      "eval_LOCATION_recall": 0.008620689655172414,
+      "eval_ORGANIZATION_f1": 0.2436974789915966,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.3918918918918919,
+      "eval_ORGANIZATION_precision": 0.3625,
       "eval_ORGANIZATION_recall": 0.18354430379746836,
-      "eval_PERSON_f1": 0.2764976958525346,
-      "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.3225806451612903,
-      "eval_PERSON_recall": 0.24193548387096775,
-      "eval_loss": 0.3704892694950104,
-      "eval_overall_accuracy": 0.8709854515509196,
-      "eval_overall_f1": 0.21070811744386875,
-      "eval_overall_precision": 0.3370165745856354,
-      "eval_overall_recall": 0.15326633165829145,
-      "eval_runtime": 0.6518,
-      "eval_samples_per_second": 260.796,
-      "eval_steps_per_second": 4.602,
+      "eval_PERSON_f1": 0.3214285714285714,
+      "eval_PERSON_number": 124,
+      "eval_PERSON_precision": 0.36,
+      "eval_PERSON_recall": 0.2903225806451613,
+      "eval_loss": 0.37283432483673096,
+      "eval_overall_accuracy": 0.8718089486686796,
+      "eval_overall_f1": 0.22372881355932206,
+      "eval_overall_precision": 0.34375,
+      "eval_overall_recall": 0.1658291457286432,
+      "eval_runtime": 0.2786,
+      "eval_samples_per_second": 610.242,
+      "eval_steps_per_second": 10.769,
       "step": 288
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.2730246782302856,
+      "grad_norm": 1.4870269298553467,
       "learning_rate": 4.8e-05,
-      "loss": 0.353,
+      "loss": 0.3537,
       "step": 384
     },
     {
       "epoch": 4.0,
-      "eval_LOCATION_f1": 0.34196891191709844,
-      "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.42857142857142855,
-      "eval_LOCATION_recall": 0.28448275862068967,
-      "eval_ORGANIZATION_f1": 0.5100286532951289,
-      "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.46596858638743455,
-      "eval_ORGANIZATION_recall": 0.5632911392405063,
-      "eval_PERSON_f1": 0.6068965517241379,
-      "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.5301204819277109,
-      "eval_PERSON_recall": 0.7096774193548387,
-      "eval_loss": 0.25001201033592224,
-      "eval_overall_accuracy": 0.9214932747735383,
-      "eval_overall_f1": 0.5048076923076923,
-      "eval_overall_precision": 0.4838709677419355,
-      "eval_overall_recall": 0.5276381909547738,
-      "eval_runtime": 0.6562,
-      "eval_samples_per_second": 259.071,
-      "eval_steps_per_second": 4.572,
+      "eval_LOCATION_f1": 0.3125,
+      "eval_LOCATION_number": 116,
+      "eval_LOCATION_precision": 0.39473684210526316,
+      "eval_LOCATION_recall": 0.25862068965517243,
+      "eval_ORGANIZATION_f1": 0.5120481927710843,
+      "eval_ORGANIZATION_number": 158,
+      "eval_ORGANIZATION_precision": 0.4885057471264368,
+      "eval_ORGANIZATION_recall": 0.5379746835443038,
+      "eval_PERSON_f1": 0.627177700348432,
+      "eval_PERSON_number": 124,
+      "eval_PERSON_precision": 0.5521472392638037,
+      "eval_PERSON_recall": 0.7258064516129032,
+      "eval_loss": 0.2518472671508789,
+      "eval_overall_accuracy": 0.9198462805380181,
+      "eval_overall_f1": 0.5055487053020961,
+      "eval_overall_precision": 0.4963680387409201,
+      "eval_overall_recall": 0.5150753768844221,
+      "eval_runtime": 0.2923,
+      "eval_samples_per_second": 581.551,
+      "eval_steps_per_second": 10.263,
       "step": 384
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1.1444727182388306,
+      "grad_norm": 1.199317455291748,
       "learning_rate": 4.75e-05,
-      "loss": 0.2551,
+      "loss": 0.2513,
       "step": 480
     },
     {
       "epoch": 5.0,
-      "eval_LOCATION_f1": 0.5650224215246636,
-      "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.5887850467289719,
-      "eval_LOCATION_recall": 0.5431034482758621,
-      "eval_ORGANIZATION_f1": 0.6628895184135977,
-      "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.6,
-      "eval_ORGANIZATION_recall": 0.740506329113924,
-      "eval_PERSON_f1": 0.8625954198473281,
-      "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.8188405797101449,
-      "eval_PERSON_recall": 0.9112903225806451,
-      "eval_loss": 0.18147629499435425,
-      "eval_overall_accuracy": 0.9516881690914082,
-      "eval_overall_f1": 0.6992840095465395,
-      "eval_overall_precision": 0.6659090909090909,
-      "eval_overall_recall": 0.7361809045226131,
-      "eval_runtime": 0.6356,
-      "eval_samples_per_second": 267.464,
-      "eval_steps_per_second": 4.72,
+      "eval_LOCATION_f1": 0.5892857142857143,
+      "eval_LOCATION_number": 116,
+      "eval_LOCATION_precision": 0.6111111111111112,
+      "eval_LOCATION_recall": 0.5689655172413793,
+      "eval_ORGANIZATION_f1": 0.6590909090909092,
+      "eval_ORGANIZATION_number": 158,
+      "eval_ORGANIZATION_precision": 0.5979381443298969,
+      "eval_ORGANIZATION_recall": 0.7341772151898734,
+      "eval_PERSON_f1": 0.8571428571428572,
+      "eval_PERSON_number": 124,
+      "eval_PERSON_precision": 0.8028169014084507,
+      "eval_PERSON_recall": 0.9193548387096774,
+      "eval_loss": 0.1811506599187851,
+      "eval_overall_accuracy": 0.9497666758166347,
+      "eval_overall_f1": 0.7030878859857482,
+      "eval_overall_precision": 0.6666666666666666,
+      "eval_overall_recall": 0.7437185929648241,
+      "eval_runtime": 0.2823,
+      "eval_samples_per_second": 602.152,
+      "eval_steps_per_second": 10.626,
       "step": 480
     },
     {
       "epoch": 6.0,
-      "grad_norm": 0.9454153776168823,
+      "grad_norm": 0.7906554341316223,
       "learning_rate": 4.7e-05,
-      "loss": 0.2031,
+      "loss": 0.1948,
       "step": 576
     },
     {
       "epoch": 6.0,
-      "eval_LOCATION_f1": 0.7866108786610879,
+      "eval_LOCATION_f1": 0.759493670886076,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.7642276422764228,
-      "eval_LOCATION_recall": 0.8103448275862069,
-      "eval_ORGANIZATION_f1": 0.7678571428571429,
+      "eval_LOCATION_precision": 0.743801652892562,
+      "eval_LOCATION_recall": 0.7758620689655172,
+      "eval_ORGANIZATION_f1": 0.7659574468085106,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.7247191011235955,
-      "eval_ORGANIZATION_recall": 0.8164556962025317,
-      "eval_PERSON_f1": 0.9420849420849421,
+      "eval_ORGANIZATION_precision": 0.7368421052631579,
+      "eval_ORGANIZATION_recall": 0.7974683544303798,
+      "eval_PERSON_f1": 0.9348659003831418,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9037037037037037,
+      "eval_PERSON_precision": 0.8905109489051095,
       "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.14112454652786255,
-      "eval_overall_accuracy": 0.9670601152895965,
-      "eval_overall_f1": 0.827338129496403,
-      "eval_overall_precision": 0.7912844036697247,
-      "eval_overall_recall": 0.8668341708542714,
-      "eval_runtime": 0.6226,
-      "eval_samples_per_second": 273.047,
-      "eval_steps_per_second": 4.818,
+      "eval_loss": 0.13590684533119202,
+      "eval_overall_accuracy": 0.9656876200933296,
+      "eval_overall_f1": 0.8174123337363967,
+      "eval_overall_precision": 0.7878787878787878,
+      "eval_overall_recall": 0.8492462311557789,
+      "eval_runtime": 0.2829,
+      "eval_samples_per_second": 600.915,
+      "eval_steps_per_second": 10.604,
       "step": 576
     },
     {
       "epoch": 7.0,
-      "grad_norm": 1.3707348108291626,
+      "grad_norm": 1.447137475013733,
       "learning_rate": 4.6500000000000005e-05,
-      "loss": 0.1672,
+      "loss": 0.1623,
       "step": 672
     },
     {
       "epoch": 7.0,
-      "eval_LOCATION_f1": 0.8166666666666667,
+      "eval_LOCATION_f1": 0.8050847457627119,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.7903225806451613,
-      "eval_LOCATION_recall": 0.8448275862068966,
-      "eval_ORGANIZATION_f1": 0.7914110429447854,
+      "eval_LOCATION_precision": 0.7916666666666666,
+      "eval_LOCATION_recall": 0.8189655172413793,
+      "eval_ORGANIZATION_f1": 0.7852760736196319,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.7678571428571429,
-      "eval_ORGANIZATION_recall": 0.8164556962025317,
-      "eval_PERSON_f1": 0.9312977099236642,
+      "eval_ORGANIZATION_precision": 0.7619047619047619,
+      "eval_ORGANIZATION_recall": 0.810126582278481,
+      "eval_PERSON_f1": 0.9457364341085271,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.8840579710144928,
+      "eval_PERSON_precision": 0.9104477611940298,
       "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.11533728241920471,
-      "eval_overall_accuracy": 0.9695306066428767,
-      "eval_overall_f1": 0.8429951690821256,
-      "eval_overall_precision": 0.8116279069767441,
-      "eval_overall_recall": 0.8768844221105527,
-      "eval_runtime": 0.6081,
-      "eval_samples_per_second": 279.559,
-      "eval_steps_per_second": 4.933,
+      "eval_loss": 0.11087985336780548,
+      "eval_overall_accuracy": 0.9700796047213834,
+      "eval_overall_f1": 0.8414634146341463,
+      "eval_overall_precision": 0.8175355450236966,
+      "eval_overall_recall": 0.8668341708542714,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.903,
+      "eval_steps_per_second": 10.692,
       "step": 672
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.7069289684295654,
+      "grad_norm": 1.1691491603851318,
       "learning_rate": 4.600000000000001e-05,
-      "loss": 0.1481,
+      "loss": 0.1397,
       "step": 768
     },
     {
       "epoch": 8.0,
-      "eval_LOCATION_f1": 0.8326530612244898,
-      "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.7906976744186046,
-      "eval_LOCATION_recall": 0.8793103448275862,
-      "eval_ORGANIZATION_f1": 0.8301886792452831,
-      "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.825,
-      "eval_ORGANIZATION_recall": 0.8354430379746836,
-      "eval_PERSON_f1": 0.9568627450980393,
-      "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9312977099236641,
-      "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.09641187638044357,
-      "eval_overall_accuracy": 0.9744715893494372,
-      "eval_overall_f1": 0.8704156479217604,
-      "eval_overall_precision": 0.8476190476190476,
-      "eval_overall_recall": 0.8944723618090452,
-      "eval_runtime": 0.6094,
-      "eval_samples_per_second": 278.959,
-      "eval_steps_per_second": 4.923,
+      "eval_LOCATION_f1": 0.8220338983050848,
+      "eval_LOCATION_number": 116,
+      "eval_LOCATION_precision": 0.8083333333333333,
+      "eval_LOCATION_recall": 0.8362068965517241,
+      "eval_ORGANIZATION_f1": 0.8220858895705523,
+      "eval_ORGANIZATION_number": 158,
+      "eval_ORGANIZATION_precision": 0.7976190476190477,
+      "eval_ORGANIZATION_recall": 0.8481012658227848,
+      "eval_PERSON_f1": 0.9647058823529411,
+      "eval_PERSON_number": 124,
+      "eval_PERSON_precision": 0.9389312977099237,
+      "eval_PERSON_recall": 0.9919354838709677,
+      "eval_loss": 0.09539666771888733,
+      "eval_overall_accuracy": 0.9739225912709305,
+      "eval_overall_f1": 0.8665850673194615,
+      "eval_overall_precision": 0.8448687350835322,
+      "eval_overall_recall": 0.8894472361809045,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.754,
+      "eval_steps_per_second": 10.69,
       "step": 768
     },
     {
       "epoch": 9.0,
-      "grad_norm": 0.9355798363685608,
+      "grad_norm": 0.8122084140777588,
       "learning_rate": 4.55e-05,
-      "loss": 0.132,
+      "loss": 0.1266,
       "step": 864
     },
     {
       "epoch": 9.0,
-      "eval_LOCATION_f1": 0.8641975308641975,
+      "eval_LOCATION_f1": 0.8559670781893004,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8267716535433071,
-      "eval_LOCATION_recall": 0.9051724137931034,
-      "eval_ORGANIZATION_f1": 0.8379204892966361,
+      "eval_LOCATION_precision": 0.8188976377952756,
+      "eval_LOCATION_recall": 0.896551724137931,
+      "eval_ORGANIZATION_f1": 0.8404907975460123,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8106508875739645,
+      "eval_ORGANIZATION_precision": 0.8154761904761905,
       "eval_ORGANIZATION_recall": 0.8670886075949367,
-      "eval_PERSON_f1": 0.9568627450980393,
+      "eval_PERSON_f1": 0.9609375,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9312977099236641,
-      "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.08645071089267731,
-      "eval_overall_accuracy": 0.9766675816634641,
+      "eval_PERSON_precision": 0.9318181818181818,
+      "eval_PERSON_recall": 0.9919354838709677,
+      "eval_loss": 0.0876610055565834,
+      "eval_overall_accuracy": 0.9761185835849574,
       "eval_overall_f1": 0.8824242424242424,
       "eval_overall_precision": 0.8524590163934426,
       "eval_overall_recall": 0.914572864321608,
-      "eval_runtime": 0.6253,
-      "eval_samples_per_second": 271.886,
-      "eval_steps_per_second": 4.798,
+      "eval_runtime": 0.2798,
+      "eval_samples_per_second": 607.653,
+      "eval_steps_per_second": 10.723,
       "step": 864
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.7637869715690613,
+      "grad_norm": 0.770068883895874,
       "learning_rate": 4.5e-05,
-      "loss": 0.119,
+      "loss": 0.1157,
       "step": 960
     },
     {
       "epoch": 10.0,
-      "eval_LOCATION_f1": 0.8666666666666666,
+      "eval_LOCATION_f1": 0.8823529411764706,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8387096774193549,
-      "eval_LOCATION_recall": 0.896551724137931,
-      "eval_ORGANIZATION_f1": 0.857142857142857,
+      "eval_LOCATION_precision": 0.860655737704918,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8624999999999999,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8414634146341463,
+      "eval_ORGANIZATION_precision": 0.8518518518518519,
       "eval_ORGANIZATION_recall": 0.8734177215189873,
-      "eval_PERSON_f1": 0.9685039370078741,
-      "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9461538461538461,
-      "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.07701876759529114,
-      "eval_overall_accuracy": 0.978040076859731,
-      "eval_overall_f1": 0.8946078431372548,
-      "eval_overall_precision": 0.8732057416267942,
-      "eval_overall_recall": 0.9170854271356784,
-      "eval_runtime": 0.625,
-      "eval_samples_per_second": 271.983,
-      "eval_steps_per_second": 4.8,
+      "eval_PERSON_f1": 0.9761904761904763,
+      "eval_PERSON_number": 124,
+      "eval_PERSON_precision": 0.9609375,
+      "eval_PERSON_recall": 0.9919354838709677,
+      "eval_loss": 0.07306713610887527,
+      "eval_overall_accuracy": 0.9799615701345046,
+      "eval_overall_f1": 0.9037037037037038,
+      "eval_overall_precision": 0.8883495145631068,
+      "eval_overall_recall": 0.9195979899497487,
+      "eval_runtime": 0.2933,
+      "eval_samples_per_second": 579.604,
+      "eval_steps_per_second": 10.228,
       "step": 960
     },
     {
       "epoch": 11.0,
-      "grad_norm": 0.7323551177978516,
+      "grad_norm": 0.5027669072151184,
       "learning_rate": 4.4500000000000004e-05,
-      "loss": 0.1117,
+      "loss": 0.1111,
       "step": 1056
     },
     {
       "epoch": 11.0,
-      "eval_LOCATION_f1": 0.8925619834710743,
+      "eval_LOCATION_f1": 0.8945147679324894,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8571428571428571,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.88125,
+      "eval_LOCATION_precision": 0.8760330578512396,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8792569659442724,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8703703703703703,
-      "eval_ORGANIZATION_recall": 0.8924050632911392,
-      "eval_PERSON_f1": 0.9761904761904763,
+      "eval_ORGANIZATION_precision": 0.8606060606060606,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
+      "eval_PERSON_f1": 0.9800796812749005,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9609375,
+      "eval_PERSON_precision": 0.968503937007874,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0722014382481575,
-      "eval_overall_accuracy": 0.9805105682130113,
-      "eval_overall_f1": 0.914004914004914,
-      "eval_overall_precision": 0.8942307692307693,
-      "eval_overall_recall": 0.9346733668341709,
-      "eval_runtime": 0.6331,
-      "eval_samples_per_second": 268.535,
-      "eval_steps_per_second": 4.739,
+      "eval_loss": 0.06731364130973816,
+      "eval_overall_accuracy": 0.9813340653307714,
+      "eval_overall_f1": 0.9149198520345253,
+      "eval_overall_precision": 0.8983050847457628,
+      "eval_overall_recall": 0.9321608040201005,
+      "eval_runtime": 0.2833,
+      "eval_samples_per_second": 600.024,
+      "eval_steps_per_second": 10.589,
       "step": 1056
     },
     {
       "epoch": 12.0,
-      "grad_norm": 0.8702762722969055,
+      "grad_norm": 0.663327693939209,
       "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.1053,
+      "loss": 0.1044,
       "step": 1152
     },
     {
       "epoch": 12.0,
-      "eval_LOCATION_f1": 0.8806584362139918,
+      "eval_LOCATION_f1": 0.8945147679324894,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.84251968503937,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.8632218844984803,
+      "eval_LOCATION_precision": 0.8760330578512396,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8765432098765433,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8304093567251462,
+      "eval_ORGANIZATION_precision": 0.8554216867469879,
       "eval_ORGANIZATION_recall": 0.8987341772151899,
-      "eval_PERSON_f1": 0.9761904761904763,
+      "eval_PERSON_f1": 0.9800796812749005,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9609375,
+      "eval_PERSON_precision": 0.968503937007874,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.07104182243347168,
-      "eval_overall_accuracy": 0.9791380730167445,
-      "eval_overall_f1": 0.9029126213592233,
-      "eval_overall_precision": 0.8732394366197183,
-      "eval_overall_recall": 0.9346733668341709,
-      "eval_runtime": 0.6246,
-      "eval_samples_per_second": 272.165,
-      "eval_steps_per_second": 4.803,
+      "eval_loss": 0.06354989856481552,
+      "eval_overall_accuracy": 0.981059566291518,
+      "eval_overall_f1": 0.9137931034482758,
+      "eval_overall_precision": 0.8961352657004831,
+      "eval_overall_recall": 0.9321608040201005,
+      "eval_runtime": 0.282,
+      "eval_samples_per_second": 602.932,
+      "eval_steps_per_second": 10.64,
       "step": 1152
     },
     {
       "epoch": 13.0,
-      "grad_norm": 4.306459426879883,
+      "grad_norm": 1.0790586471557617,
       "learning_rate": 4.35e-05,
-      "loss": 0.1015,
+      "loss": 0.098,
       "step": 1248
     },
     {
       "epoch": 13.0,
-      "eval_LOCATION_f1": 0.9059829059829059,
+      "eval_LOCATION_f1": 0.8974358974358974,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8983050847457628,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8765432098765433,
+      "eval_LOCATION_precision": 0.8898305084745762,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8722741433021807,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8554216867469879,
-      "eval_ORGANIZATION_recall": 0.8987341772151899,
-      "eval_PERSON_f1": 0.9682539682539683,
+      "eval_ORGANIZATION_precision": 0.8588957055214724,
+      "eval_ORGANIZATION_recall": 0.8860759493670886,
+      "eval_PERSON_f1": 0.9840000000000001,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.953125,
-      "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.05972462147474289,
-      "eval_overall_accuracy": 0.9818830634092781,
-      "eval_overall_f1": 0.9135802469135802,
-      "eval_overall_precision": 0.8980582524271845,
-      "eval_overall_recall": 0.9296482412060302,
-      "eval_runtime": 0.6233,
-      "eval_samples_per_second": 272.73,
-      "eval_steps_per_second": 4.813,
+      "eval_PERSON_precision": 0.9761904761904762,
+      "eval_PERSON_recall": 0.9919354838709677,
+      "eval_loss": 0.057803478091955185,
+      "eval_overall_accuracy": 0.9816085643700247,
+      "eval_overall_f1": 0.9142857142857144,
+      "eval_overall_precision": 0.9041769041769042,
+      "eval_overall_recall": 0.9246231155778895,
+      "eval_runtime": 0.2823,
+      "eval_samples_per_second": 602.275,
+      "eval_steps_per_second": 10.628,
       "step": 1248
     },
     {
       "epoch": 14.0,
-      "grad_norm": 1.0401694774627686,
+      "grad_norm": 1.317100167274475,
       "learning_rate": 4.3e-05,
-      "loss": 0.0979,
+      "loss": 0.0939,
       "step": 1344
     },
     {
       "epoch": 14.0,
-      "eval_LOCATION_f1": 0.8945147679324894,
+      "eval_LOCATION_f1": 0.8898305084745763,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8760330578512396,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8832807570977917,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8749999999999999,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8805031446540881,
+      "eval_ORGANIZATION_precision": 0.8641975308641975,
       "eval_ORGANIZATION_recall": 0.8860759493670886,
-      "eval_PERSON_f1": 0.9800796812749005,
+      "eval_PERSON_f1": 0.9840000000000001,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.968503937007874,
+      "eval_PERSON_precision": 0.9761904761904762,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.059020232409238815,
-      "eval_overall_accuracy": 0.9816085643700247,
-      "eval_overall_f1": 0.9167701863354037,
-      "eval_overall_precision": 0.9066339066339066,
-      "eval_overall_recall": 0.9271356783919598,
-      "eval_runtime": 0.6209,
-      "eval_samples_per_second": 273.815,
-      "eval_steps_per_second": 4.832,
+      "eval_loss": 0.05590749531984329,
+      "eval_overall_accuracy": 0.9818830634092781,
+      "eval_overall_f1": 0.9131513647642681,
+      "eval_overall_precision": 0.9019607843137255,
+      "eval_overall_recall": 0.9246231155778895,
+      "eval_runtime": 0.2821,
+      "eval_samples_per_second": 602.62,
+      "eval_steps_per_second": 10.634,
       "step": 1344
     },
     {
       "epoch": 15.0,
-      "grad_norm": 0.8894268274307251,
+      "grad_norm": 0.728569507598877,
       "learning_rate": 4.25e-05,
-      "loss": 0.0929,
+      "loss": 0.091,
       "step": 1440
     },
     {
       "epoch": 15.0,
-      "eval_LOCATION_f1": 0.8983050847457628,
+      "eval_LOCATION_f1": 0.8936170212765958,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8833333333333333,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8765432098765433,
+      "eval_LOCATION_precision": 0.8823529411764706,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8685015290519877,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8554216867469879,
+      "eval_ORGANIZATION_precision": 0.8402366863905325,
       "eval_ORGANIZATION_recall": 0.8987341772151899,
-      "eval_PERSON_f1": 0.9761904761904763,
+      "eval_PERSON_f1": 0.9800796812749005,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9609375,
+      "eval_PERSON_precision": 0.968503937007874,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.05802612379193306,
-      "eval_overall_accuracy": 0.981059566291518,
-      "eval_overall_f1": 0.9137931034482758,
-      "eval_overall_precision": 0.8961352657004831,
-      "eval_overall_recall": 0.9321608040201005,
-      "eval_runtime": 0.626,
-      "eval_samples_per_second": 271.549,
-      "eval_steps_per_second": 4.792,
+      "eval_loss": 0.05577402561903,
+      "eval_overall_accuracy": 0.9816085643700247,
+      "eval_overall_f1": 0.910209102091021,
+      "eval_overall_precision": 0.891566265060241,
+      "eval_overall_recall": 0.9296482412060302,
+      "eval_runtime": 0.2813,
+      "eval_samples_per_second": 604.331,
+      "eval_steps_per_second": 10.665,
       "step": 1440
     },
     {
       "epoch": 16.0,
-      "grad_norm": 1.4922378063201904,
+      "grad_norm": 1.4055036306381226,
       "learning_rate": 4.2e-05,
-      "loss": 0.0881,
+      "loss": 0.088,
       "step": 1536
     },
     {
       "epoch": 16.0,
-      "eval_LOCATION_f1": 0.8945147679324894,
+      "eval_LOCATION_f1": 0.8898305084745763,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8760330578512396,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8545454545454545,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8711656441717791,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8197674418604651,
-      "eval_ORGANIZATION_recall": 0.8924050632911392,
-      "eval_PERSON_f1": 0.9761904761904763,
+      "eval_ORGANIZATION_precision": 0.8452380952380952,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
+      "eval_PERSON_f1": 0.9723320158102766,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9609375,
+      "eval_PERSON_precision": 0.9534883720930233,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.058439191430807114,
-      "eval_overall_accuracy": 0.9802360691737579,
-      "eval_overall_f1": 0.9035409035409036,
-      "eval_overall_precision": 0.8788598574821853,
+      "eval_loss": 0.05548899993300438,
+      "eval_overall_accuracy": 0.981059566291518,
+      "eval_overall_f1": 0.9079754601226994,
+      "eval_overall_precision": 0.8872901678657075,
       "eval_overall_recall": 0.9296482412060302,
-      "eval_runtime": 0.6225,
-      "eval_samples_per_second": 273.096,
-      "eval_steps_per_second": 4.819,
+      "eval_runtime": 0.2808,
+      "eval_samples_per_second": 605.486,
+      "eval_steps_per_second": 10.685,
       "step": 1536
     },
     {
       "epoch": 17.0,
-      "grad_norm": 0.8413049578666687,
+      "grad_norm": 0.5631800293922424,
       "learning_rate": 4.15e-05,
-      "loss": 0.0849,
+      "loss": 0.0857,
       "step": 1632
     },
     {
       "epoch": 17.0,
-      "eval_LOCATION_f1": 0.8813559322033899,
+      "eval_LOCATION_f1": 0.8936170212765958,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8666666666666667,
-      "eval_LOCATION_recall": 0.896551724137931,
-      "eval_ORGANIZATION_f1": 0.8805031446540881,
+      "eval_LOCATION_precision": 0.8823529411764706,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.889589905362776,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.875,
-      "eval_ORGANIZATION_recall": 0.8860759493670886,
-      "eval_PERSON_f1": 0.9840000000000001,
+      "eval_ORGANIZATION_precision": 0.8867924528301887,
+      "eval_ORGANIZATION_recall": 0.8924050632911392,
+      "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9761904761904762,
+      "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.05292287841439247,
-      "eval_overall_accuracy": 0.9818830634092781,
-      "eval_overall_f1": 0.9129353233830846,
-      "eval_overall_precision": 0.9039408866995073,
-      "eval_overall_recall": 0.9221105527638191,
-      "eval_runtime": 0.623,
-      "eval_samples_per_second": 272.855,
-      "eval_steps_per_second": 4.815,
+      "eval_loss": 0.052291139960289,
+      "eval_overall_accuracy": 0.9846280538018117,
+      "eval_overall_f1": 0.9213483146067416,
+      "eval_overall_precision": 0.9156327543424317,
+      "eval_overall_recall": 0.9271356783919598,
+      "eval_runtime": 0.2834,
+      "eval_samples_per_second": 599.885,
+      "eval_steps_per_second": 10.586,
       "step": 1632
     },
     {
       "epoch": 18.0,
-      "grad_norm": 0.9637477993965149,
+      "grad_norm": 0.8045436143875122,
       "learning_rate": 4.1e-05,
-      "loss": 0.0843,
+      "loss": 0.0809,
       "step": 1728
     },
     {
       "epoch": 18.0,
-      "eval_LOCATION_f1": 0.9205020920502092,
+      "eval_LOCATION_f1": 0.8860759493670886,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8943089430894309,
-      "eval_LOCATION_recall": 0.9482758620689655,
-      "eval_ORGANIZATION_f1": 0.8785046728971962,
+      "eval_LOCATION_precision": 0.8677685950413223,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8819875776397516,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8650306748466258,
-      "eval_ORGANIZATION_recall": 0.8924050632911392,
-      "eval_PERSON_f1": 0.9800796812749005,
+      "eval_ORGANIZATION_precision": 0.8658536585365854,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
+      "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.968503937007874,
+      "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.05260878801345825,
-      "eval_overall_accuracy": 0.9829810595662916,
-      "eval_overall_f1": 0.9223181257706535,
-      "eval_overall_precision": 0.9055690072639225,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.616,
-      "eval_samples_per_second": 275.955,
-      "eval_steps_per_second": 4.87,
+      "eval_loss": 0.04978971555829048,
+      "eval_overall_accuracy": 0.9832555586055449,
+      "eval_overall_f1": 0.9158415841584159,
+      "eval_overall_precision": 0.9024390243902439,
+      "eval_overall_recall": 0.9296482412060302,
+      "eval_runtime": 0.2824,
+      "eval_samples_per_second": 602.083,
+      "eval_steps_per_second": 10.625,
       "step": 1728
     },
     {
       "epoch": 19.0,
-      "grad_norm": 0.7680073976516724,
+      "grad_norm": 0.8581104278564453,
       "learning_rate": 4.05e-05,
-      "loss": 0.0826,
+      "loss": 0.0773,
       "step": 1824
     },
     {
       "epoch": 19.0,
-      "eval_LOCATION_f1": 0.9059829059829059,
+      "eval_LOCATION_f1": 0.8974358974358974,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8983050847457628,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8847352024922118,
+      "eval_LOCATION_precision": 0.8898305084745762,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.89375,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8711656441717791,
-      "eval_ORGANIZATION_recall": 0.8987341772151899,
+      "eval_ORGANIZATION_precision": 0.8827160493827161,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04939362034201622,
-      "eval_overall_accuracy": 0.9835300576447983,
-      "eval_overall_f1": 0.9228855721393034,
-      "eval_overall_precision": 0.9137931034482759,
+      "eval_loss": 0.048203177750110626,
+      "eval_overall_accuracy": 0.9843535547625584,
+      "eval_overall_f1": 0.9240348692403487,
+      "eval_overall_precision": 0.9160493827160494,
       "eval_overall_recall": 0.9321608040201005,
-      "eval_runtime": 0.6255,
-      "eval_samples_per_second": 271.772,
-      "eval_steps_per_second": 4.796,
+      "eval_runtime": 0.281,
+      "eval_samples_per_second": 604.968,
+      "eval_steps_per_second": 10.676,
       "step": 1824
     },
     {
       "epoch": 20.0,
-      "grad_norm": 0.9873287081718445,
+      "grad_norm": 0.8157017827033997,
       "learning_rate": 4e-05,
-      "loss": 0.0775,
+      "loss": 0.0765,
       "step": 1920
     },
     {
       "epoch": 20.0,
-      "eval_LOCATION_f1": 0.903765690376569,
+      "eval_LOCATION_f1": 0.8983050847457628,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8780487804878049,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.88,
+      "eval_LOCATION_precision": 0.8833333333333333,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8834355828220858,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8562874251497006,
-      "eval_ORGANIZATION_recall": 0.9050632911392406,
-      "eval_PERSON_f1": 0.9840000000000001,
+      "eval_ORGANIZATION_precision": 0.8571428571428571,
+      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_PERSON_f1": 0.9800796812749005,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9761904761904762,
+      "eval_PERSON_precision": 0.968503937007874,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.050805024802684784,
+      "eval_loss": 0.05213193967938423,
       "eval_overall_accuracy": 0.9821575624485315,
-      "eval_overall_f1": 0.918918918918919,
-      "eval_overall_precision": 0.8990384615384616,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6257,
-      "eval_samples_per_second": 271.682,
-      "eval_steps_per_second": 4.794,
+      "eval_overall_f1": 0.9175891758917589,
+      "eval_overall_precision": 0.8987951807228916,
+      "eval_overall_recall": 0.9371859296482412,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 603.469,
+      "eval_steps_per_second": 10.649,
       "step": 1920
     },
     {
       "epoch": 21.0,
-      "grad_norm": 1.1787539720535278,
+      "grad_norm": 1.5243802070617676,
       "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.0725,
+      "loss": 0.0754,
       "step": 2016
     },
     {
       "epoch": 21.0,
-      "eval_LOCATION_f1": 0.9029535864978903,
+      "eval_LOCATION_f1": 0.8898305084745763,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8842975206611571,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.8834355828220858,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8950617283950618,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8571428571428571,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.8734939759036144,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04908851161599159,
-      "eval_overall_accuracy": 0.9832555586055449,
-      "eval_overall_f1": 0.9211822660098522,
-      "eval_overall_precision": 0.9033816425120773,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.627,
-      "eval_samples_per_second": 271.141,
-      "eval_steps_per_second": 4.785,
+      "eval_loss": 0.0483800433576107,
+      "eval_overall_accuracy": 0.984079055723305,
+      "eval_overall_f1": 0.9221260815822002,
+      "eval_overall_precision": 0.9075425790754258,
+      "eval_overall_recall": 0.9371859296482412,
+      "eval_runtime": 0.2805,
+      "eval_samples_per_second": 606.104,
+      "eval_steps_per_second": 10.696,
       "step": 2016
     },
     {
       "epoch": 22.0,
-      "grad_norm": 1.0315650701522827,
+      "grad_norm": 1.5905965566635132,
       "learning_rate": 3.9000000000000006e-05,
-      "loss": 0.0712,
+      "loss": 0.072,
       "step": 2112
     },
     {
       "epoch": 22.0,
-      "eval_LOCATION_f1": 0.9059829059829059,
+      "eval_LOCATION_f1": 0.8898305084745763,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8983050847457628,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8944099378881988,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8792569659442724,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8780487804878049,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
-      "eval_PERSON_f1": 0.9840000000000001,
+      "eval_ORGANIZATION_precision": 0.8606060606060606,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
+      "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9761904761904762,
+      "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04578203707933426,
+      "eval_loss": 0.046878568828105927,
       "eval_overall_accuracy": 0.9835300576447983,
-      "eval_overall_f1": 0.9255583126550867,
-      "eval_overall_precision": 0.9142156862745098,
-      "eval_overall_recall": 0.9371859296482412,
-      "eval_runtime": 0.6301,
-      "eval_samples_per_second": 269.818,
-      "eval_steps_per_second": 4.761,
+      "eval_overall_f1": 0.9158415841584159,
+      "eval_overall_precision": 0.9024390243902439,
+      "eval_overall_recall": 0.9296482412060302,
+      "eval_runtime": 0.2799,
+      "eval_samples_per_second": 607.427,
+      "eval_steps_per_second": 10.719,
       "step": 2112
     },
     {
       "epoch": 23.0,
-      "grad_norm": 0.9216381311416626,
+      "grad_norm": 0.7514567375183105,
       "learning_rate": 3.85e-05,
-      "loss": 0.071,
+      "loss": 0.0689,
       "step": 2208
     },
     {
       "epoch": 23.0,
-      "eval_LOCATION_f1": 0.8907563025210085,
+      "eval_LOCATION_f1": 0.8974358974358974,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8688524590163934,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.889589905362776,
+      "eval_LOCATION_precision": 0.8898305084745762,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.9028213166144201,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8867924528301887,
-      "eval_ORGANIZATION_recall": 0.8924050632911392,
+      "eval_ORGANIZATION_precision": 0.8944099378881988,
+      "eval_ORGANIZATION_recall": 0.9113924050632911,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.045151468366384506,
-      "eval_overall_accuracy": 0.984079055723305,
-      "eval_overall_f1": 0.9203980099502488,
-      "eval_overall_precision": 0.9113300492610837,
-      "eval_overall_recall": 0.9296482412060302,
-      "eval_runtime": 0.6299,
-      "eval_samples_per_second": 269.88,
-      "eval_steps_per_second": 4.763,
+      "eval_loss": 0.044030942022800446,
+      "eval_overall_accuracy": 0.9843535547625584,
+      "eval_overall_f1": 0.9276807980049875,
+      "eval_overall_precision": 0.9207920792079208,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.822,
+      "eval_steps_per_second": 10.691,
       "step": 2208
     },
     {
       "epoch": 24.0,
-      "grad_norm": 2.3887767791748047,
+      "grad_norm": 1.4454855918884277,
       "learning_rate": 3.8e-05,
-      "loss": 0.0708,
+      "loss": 0.0697,
       "step": 2304
     },
     {
       "epoch": 24.0,
-      "eval_LOCATION_f1": 0.8907563025210085,
+      "eval_LOCATION_f1": 0.9012875536480686,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8688524590163934,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.889589905362776,
+      "eval_LOCATION_precision": 0.8974358974358975,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.8881789137380192,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8867924528301887,
-      "eval_ORGANIZATION_recall": 0.8924050632911392,
+      "eval_ORGANIZATION_precision": 0.896774193548387,
+      "eval_ORGANIZATION_recall": 0.879746835443038,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04399710148572922,
-      "eval_overall_accuracy": 0.9849025528410651,
-      "eval_overall_f1": 0.9203980099502488,
-      "eval_overall_precision": 0.9113300492610837,
-      "eval_overall_recall": 0.9296482412060302,
-      "eval_runtime": 0.629,
-      "eval_samples_per_second": 270.279,
-      "eval_steps_per_second": 4.77,
+      "eval_loss": 0.04559732973575592,
+      "eval_overall_accuracy": 0.9846280538018117,
+      "eval_overall_f1": 0.9232704402515725,
+      "eval_overall_precision": 0.924433249370277,
+      "eval_overall_recall": 0.9221105527638191,
+      "eval_runtime": 0.2825,
+      "eval_samples_per_second": 601.735,
+      "eval_steps_per_second": 10.619,
       "step": 2304
     },
     {
       "epoch": 25.0,
-      "grad_norm": 0.9052558541297913,
+      "grad_norm": 1.2895970344543457,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0681,
+      "loss": 0.0656,
       "step": 2400
     },
     {
       "epoch": 25.0,
-      "eval_LOCATION_f1": 0.8983050847457628,
+      "eval_LOCATION_f1": 0.9059829059829059,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8833333333333333,
+      "eval_LOCATION_precision": 0.8983050847457628,
       "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.8999999999999999,
+      "eval_ORGANIZATION_f1": 0.8867924528301886,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8888888888888888,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.88125,
+      "eval_ORGANIZATION_recall": 0.8924050632911392,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0418149009346962,
-      "eval_overall_accuracy": 0.9854515509195718,
-      "eval_overall_f1": 0.9267080745341615,
-      "eval_overall_precision": 0.9164619164619164,
-      "eval_overall_recall": 0.9371859296482412,
-      "eval_runtime": 0.6322,
-      "eval_samples_per_second": 268.906,
-      "eval_steps_per_second": 4.745,
+      "eval_loss": 0.04359252005815506,
+      "eval_overall_accuracy": 0.9846280538018117,
+      "eval_overall_f1": 0.9238451935081149,
+      "eval_overall_precision": 0.9181141439205955,
+      "eval_overall_recall": 0.9296482412060302,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 603.476,
+      "eval_steps_per_second": 10.65,
       "step": 2400
     },
     {
       "epoch": 26.0,
-      "grad_norm": 0.3684695065021515,
+      "grad_norm": 0.6096397638320923,
       "learning_rate": 3.7e-05,
-      "loss": 0.0637,
+      "loss": 0.0658,
       "step": 2496
     },
     {
       "epoch": 26.0,
-      "eval_LOCATION_f1": 0.9106382978723405,
+      "eval_LOCATION_f1": 0.9012875536480686,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8991596638655462,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.897196261682243,
+      "eval_LOCATION_precision": 0.8974358974358975,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.88125,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8834355828220859,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.8703703703703703,
+      "eval_ORGANIZATION_recall": 0.8924050632911392,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.041572462767362595,
-      "eval_overall_accuracy": 0.9860005489980785,
-      "eval_overall_f1": 0.9291925465838509,
-      "eval_overall_precision": 0.918918918918919,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6288,
-      "eval_samples_per_second": 270.365,
-      "eval_steps_per_second": 4.771,
+      "eval_loss": 0.0426979586482048,
+      "eval_overall_accuracy": 0.984079055723305,
+      "eval_overall_f1": 0.9201995012468827,
+      "eval_overall_precision": 0.9133663366336634,
+      "eval_overall_recall": 0.9271356783919598,
+      "eval_runtime": 0.2821,
+      "eval_samples_per_second": 602.528,
+      "eval_steps_per_second": 10.633,
       "step": 2496
     },
     {
       "epoch": 27.0,
-      "grad_norm": 1.3911359310150146,
+      "grad_norm": 1.2615808248519897,
       "learning_rate": 3.65e-05,
-      "loss": 0.0644,
+      "loss": 0.065,
       "step": 2592
     },
     {
       "epoch": 27.0,
-      "eval_LOCATION_f1": 0.9067796610169492,
+      "eval_LOCATION_f1": 0.9051724137931035,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.8916666666666667,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.8978328173374612,
+      "eval_LOCATION_precision": 0.9051724137931034,
+      "eval_LOCATION_recall": 0.9051724137931034,
+      "eval_ORGANIZATION_f1": 0.897196261682243,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8787878787878788,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.8834355828220859,
+      "eval_ORGANIZATION_recall": 0.9113924050632911,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04298856854438782,
-      "eval_overall_accuracy": 0.9860005489980785,
-      "eval_overall_f1": 0.9282178217821783,
-      "eval_overall_precision": 0.9146341463414634,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6314,
-      "eval_samples_per_second": 269.228,
-      "eval_steps_per_second": 4.751,
+      "eval_loss": 0.04214628413319588,
+      "eval_overall_accuracy": 0.9854515509195718,
+      "eval_overall_f1": 0.9276807980049875,
+      "eval_overall_precision": 0.9207920792079208,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2802,
+      "eval_samples_per_second": 606.749,
+      "eval_steps_per_second": 10.707,
       "step": 2592
     },
     {
       "epoch": 28.0,
-      "grad_norm": 1.201261281967163,
+      "grad_norm": 1.0637108087539673,
       "learning_rate": 3.6e-05,
-      "loss": 0.0617,
+      "loss": 0.0613,
       "step": 2688
     },
     {
       "epoch": 28.0,
-      "eval_LOCATION_f1": 0.9191489361702128,
+      "eval_LOCATION_f1": 0.8983050847457628,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.907563025210084,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9040247678018577,
+      "eval_LOCATION_precision": 0.8833333333333333,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8965517241379312,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8848484848484849,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.8881987577639752,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.042760126292705536,
-      "eval_overall_accuracy": 0.9865495470765853,
-      "eval_overall_f1": 0.9343246592317224,
-      "eval_overall_precision": 0.921760391198044,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.627,
-      "eval_samples_per_second": 271.117,
-      "eval_steps_per_second": 4.784,
+      "eval_loss": 0.041786737740039825,
+      "eval_overall_accuracy": 0.9854515509195718,
+      "eval_overall_f1": 0.9253731343283583,
+      "eval_overall_precision": 0.916256157635468,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.281,
+      "eval_samples_per_second": 604.884,
+      "eval_steps_per_second": 10.674,
       "step": 2688
     },
     {
       "epoch": 29.0,
-      "grad_norm": 0.6504961848258972,
+      "grad_norm": 0.4293781816959381,
       "learning_rate": 3.55e-05,
-      "loss": 0.0615,
+      "loss": 0.0591,
       "step": 2784
     },
     {
       "epoch": 29.0,
-      "eval_LOCATION_f1": 0.923076923076923,
+      "eval_LOCATION_f1": 0.9098712446351932,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9152542372881356,
-      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.905982905982906,
+      "eval_LOCATION_recall": 0.9137931034482759,
       "eval_ORGANIZATION_f1": 0.8965517241379312,
       "eval_ORGANIZATION_number": 158,
       "eval_ORGANIZATION_precision": 0.8881987577639752,
@@ -897,1013 +897,1013 @@
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04017549008131027,
-      "eval_overall_accuracy": 0.9862750480373319,
-      "eval_overall_f1": 0.9326683291770573,
-      "eval_overall_precision": 0.9257425742574258,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6891,
-      "eval_samples_per_second": 246.703,
-      "eval_steps_per_second": 4.354,
+      "eval_loss": 0.039750631898641586,
+      "eval_overall_accuracy": 0.9873730441943454,
+      "eval_overall_f1": 0.9288389513108615,
+      "eval_overall_precision": 0.9230769230769231,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2819,
+      "eval_samples_per_second": 603.055,
+      "eval_steps_per_second": 10.642,
       "step": 2784
     },
     {
       "epoch": 30.0,
-      "grad_norm": 1.9796297550201416,
+      "grad_norm": 1.076916217803955,
       "learning_rate": 3.5e-05,
-      "loss": 0.055,
+      "loss": 0.06,
       "step": 2880
     },
     {
       "epoch": 30.0,
-      "eval_LOCATION_f1": 0.9184549356223175,
+      "eval_LOCATION_f1": 0.9098712446351932,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9145299145299145,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9068322981366461,
+      "eval_LOCATION_precision": 0.905982905982906,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.9022082018927445,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8902439024390244,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.89937106918239,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03879169002175331,
-      "eval_overall_accuracy": 0.9868240461158386,
-      "eval_overall_f1": 0.9353233830845771,
-      "eval_overall_precision": 0.9261083743842364,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6318,
-      "eval_samples_per_second": 269.066,
-      "eval_steps_per_second": 4.748,
+      "eval_loss": 0.03954719007015228,
+      "eval_overall_accuracy": 0.9865495470765853,
+      "eval_overall_f1": 0.9311639549436797,
+      "eval_overall_precision": 0.9276807980049875,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2821,
+      "eval_samples_per_second": 602.55,
+      "eval_steps_per_second": 10.633,
       "step": 2880
     },
     {
       "epoch": 31.0,
-      "grad_norm": 0.7729928493499756,
+      "grad_norm": 0.6545688509941101,
       "learning_rate": 3.45e-05,
-      "loss": 0.0582,
+      "loss": 0.0566,
       "step": 2976
     },
     {
       "epoch": 31.0,
-      "eval_LOCATION_f1": 0.9191489361702128,
+      "eval_LOCATION_f1": 0.9059829059829059,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.907563025210084,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9107692307692307,
+      "eval_LOCATION_precision": 0.8983050847457628,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.89375,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8862275449101796,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.8827160493827161,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03867227956652641,
-      "eval_overall_accuracy": 0.9873730441943454,
-      "eval_overall_f1": 0.9369592088998764,
-      "eval_overall_precision": 0.9221411192214112,
-      "eval_overall_recall": 0.9522613065326633,
-      "eval_runtime": 0.6223,
-      "eval_samples_per_second": 273.165,
-      "eval_steps_per_second": 4.821,
+      "eval_loss": 0.03855058550834656,
+      "eval_overall_accuracy": 0.9862750480373319,
+      "eval_overall_f1": 0.9265255292652553,
+      "eval_overall_precision": 0.9185185185185185,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.279,
+      "eval_samples_per_second": 609.398,
+      "eval_steps_per_second": 10.754,
       "step": 2976
     },
     {
       "epoch": 32.0,
-      "grad_norm": 1.2700538635253906,
+      "grad_norm": 1.2271991968154907,
       "learning_rate": 3.4000000000000007e-05,
-      "loss": 0.0567,
+      "loss": 0.0566,
       "step": 3072
     },
     {
       "epoch": 32.0,
-      "eval_LOCATION_f1": 0.9191489361702128,
+      "eval_LOCATION_f1": 0.8927038626609441,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.907563025210084,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.8999999999999999,
+      "eval_LOCATION_precision": 0.8888888888888888,
+      "eval_LOCATION_recall": 0.896551724137931,
+      "eval_ORGANIZATION_f1": 0.9015873015873016,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8888888888888888,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.9044585987261147,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03720666840672493,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9328358208955224,
-      "eval_overall_precision": 0.9236453201970444,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6263,
-      "eval_samples_per_second": 271.437,
-      "eval_steps_per_second": 4.79,
+      "eval_loss": 0.03917478770017624,
+      "eval_overall_accuracy": 0.9857260499588252,
+      "eval_overall_f1": 0.9259723964868257,
+      "eval_overall_precision": 0.924812030075188,
+      "eval_overall_recall": 0.9271356783919598,
+      "eval_runtime": 0.2818,
+      "eval_samples_per_second": 603.321,
+      "eval_steps_per_second": 10.647,
       "step": 3072
     },
     {
       "epoch": 33.0,
-      "grad_norm": 0.7994013428688049,
+      "grad_norm": 0.7312807440757751,
       "learning_rate": 3.35e-05,
-      "loss": 0.0581,
+      "loss": 0.0566,
       "step": 3168
     },
     {
       "epoch": 33.0,
-      "eval_LOCATION_f1": 0.9152542372881356,
+      "eval_LOCATION_f1": 0.9106382978723405,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9068322981366461,
+      "eval_LOCATION_precision": 0.8991596638655462,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.9015873015873016,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.8902439024390244,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9044585987261147,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.038004301488399506,
-      "eval_overall_accuracy": 0.9876475432335987,
-      "eval_overall_f1": 0.9343246592317224,
-      "eval_overall_precision": 0.921760391198044,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6369,
-      "eval_samples_per_second": 266.898,
-      "eval_steps_per_second": 4.71,
+      "eval_loss": 0.039775025099515915,
+      "eval_overall_accuracy": 0.9865495470765853,
+      "eval_overall_f1": 0.9311639549436797,
+      "eval_overall_precision": 0.9276807980049875,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2816,
+      "eval_samples_per_second": 603.775,
+      "eval_steps_per_second": 10.655,
       "step": 3168
     },
     {
       "epoch": 34.0,
-      "grad_norm": 1.298720121383667,
+      "grad_norm": 0.8602070212364197,
       "learning_rate": 3.3e-05,
-      "loss": 0.0534,
+      "loss": 0.0568,
       "step": 3264
     },
     {
       "epoch": 34.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9224137931034483,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9254658385093167,
+      "eval_LOCATION_precision": 0.9224137931034483,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.90625,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9085365853658537,
-      "eval_ORGANIZATION_recall": 0.9430379746835443,
+      "eval_ORGANIZATION_precision": 0.8950617283950617,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.039352480322122574,
-      "eval_overall_accuracy": 0.9876475432335987,
-      "eval_overall_f1": 0.946450809464508,
-      "eval_overall_precision": 0.9382716049382716,
-      "eval_overall_recall": 0.9547738693467337,
-      "eval_runtime": 0.6236,
-      "eval_samples_per_second": 272.632,
-      "eval_steps_per_second": 4.811,
+      "eval_loss": 0.03961801528930664,
+      "eval_overall_accuracy": 0.987098545155092,
+      "eval_overall_f1": 0.9363295880149813,
+      "eval_overall_precision": 0.9305210918114144,
+      "eval_overall_recall": 0.9422110552763819,
+      "eval_runtime": 0.2818,
+      "eval_samples_per_second": 603.331,
+      "eval_steps_per_second": 10.647,
       "step": 3264
     },
     {
       "epoch": 35.0,
-      "grad_norm": 1.8338191509246826,
+      "grad_norm": 1.1235308647155762,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.0536,
+      "loss": 0.0532,
       "step": 3360
     },
     {
       "epoch": 35.0,
-      "eval_LOCATION_f1": 0.923076923076923,
+      "eval_LOCATION_f1": 0.9059829059829059,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9152542372881356,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9254658385093167,
+      "eval_LOCATION_precision": 0.8983050847457628,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.9050632911392406,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9085365853658537,
-      "eval_ORGANIZATION_recall": 0.9430379746835443,
+      "eval_ORGANIZATION_precision": 0.9050632911392406,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03686641529202461,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9440993788819877,
-      "eval_overall_precision": 0.9336609336609336,
-      "eval_overall_recall": 0.9547738693467337,
-      "eval_runtime": 0.6916,
-      "eval_samples_per_second": 245.813,
-      "eval_steps_per_second": 4.338,
+      "eval_loss": 0.037914156913757324,
+      "eval_overall_accuracy": 0.987098545155092,
+      "eval_overall_f1": 0.9311639549436797,
+      "eval_overall_precision": 0.9276807980049875,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2789,
+      "eval_samples_per_second": 609.623,
+      "eval_steps_per_second": 10.758,
       "step": 3360
     },
     {
       "epoch": 36.0,
-      "grad_norm": 1.2448095083236694,
+      "grad_norm": 0.49202960729599,
       "learning_rate": 3.2000000000000005e-05,
-      "loss": 0.0511,
+      "loss": 0.052,
       "step": 3456
     },
     {
       "epoch": 36.0,
-      "eval_LOCATION_f1": 0.9243697478991596,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9016393442622951,
-      "eval_LOCATION_recall": 0.9482758620689655,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.9230769230769231,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9125,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
+      "eval_ORGANIZATION_precision": 0.9012345679012346,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.04009932279586792,
-      "eval_overall_accuracy": 0.9876475432335987,
-      "eval_overall_f1": 0.9427860696517413,
-      "eval_overall_precision": 0.9334975369458128,
-      "eval_overall_recall": 0.9522613065326633,
-      "eval_runtime": 0.6343,
-      "eval_samples_per_second": 268.001,
-      "eval_steps_per_second": 4.729,
+      "eval_loss": 0.04034107178449631,
+      "eval_overall_accuracy": 0.9879220422728521,
+      "eval_overall_f1": 0.9401496259351622,
+      "eval_overall_precision": 0.9331683168316832,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2807,
+      "eval_samples_per_second": 605.629,
+      "eval_steps_per_second": 10.688,
       "step": 3456
     },
     {
       "epoch": 37.0,
-      "grad_norm": 1.1896450519561768,
+      "grad_norm": 1.104040503501892,
       "learning_rate": 3.15e-05,
-      "loss": 0.0512,
+      "loss": 0.0516,
       "step": 3552
     },
     {
       "epoch": 37.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.9059829059829059,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.8983050847457628,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.9056603773584907,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9,
+      "eval_ORGANIZATION_recall": 0.9113924050632911,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.038715094327926636,
-      "eval_overall_accuracy": 0.9868240461158386,
-      "eval_overall_f1": 0.9422110552763819,
-      "eval_overall_precision": 0.9422110552763819,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6255,
-      "eval_samples_per_second": 271.794,
-      "eval_steps_per_second": 4.796,
+      "eval_loss": 0.03864772990345955,
+      "eval_overall_accuracy": 0.9873730441943454,
+      "eval_overall_f1": 0.9313358302122346,
+      "eval_overall_precision": 0.9255583126550868,
+      "eval_overall_recall": 0.9371859296482412,
+      "eval_runtime": 0.2787,
+      "eval_samples_per_second": 609.902,
+      "eval_steps_per_second": 10.763,
       "step": 3552
     },
     {
       "epoch": 38.0,
-      "grad_norm": 0.7601023316383362,
+      "grad_norm": 0.9916018843650818,
       "learning_rate": 3.1e-05,
-      "loss": 0.0498,
+      "loss": 0.0497,
       "step": 3648
     },
     {
       "epoch": 38.0,
-      "eval_LOCATION_f1": 0.9184549356223175,
+      "eval_LOCATION_f1": 0.9106382978723405,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9145299145299145,
+      "eval_LOCATION_precision": 0.8991596638655462,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9148264984227129,
+      "eval_ORGANIZATION_f1": 0.9022082018927445,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9119496855345912,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.89937106918239,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03696443885564804,
+      "eval_loss": 0.03776225075125694,
       "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9386733416770965,
-      "eval_overall_precision": 0.9351620947630923,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6226,
-      "eval_samples_per_second": 273.03,
-      "eval_steps_per_second": 4.818,
+      "eval_overall_f1": 0.9313358302122346,
+      "eval_overall_precision": 0.9255583126550868,
+      "eval_overall_recall": 0.9371859296482412,
+      "eval_runtime": 0.2818,
+      "eval_samples_per_second": 603.336,
+      "eval_steps_per_second": 10.647,
       "step": 3648
     },
     {
       "epoch": 39.0,
-      "grad_norm": 1.1390430927276611,
+      "grad_norm": 0.43910089135169983,
       "learning_rate": 3.05e-05,
-      "loss": 0.0489,
+      "loss": 0.052,
       "step": 3744
     },
     {
       "epoch": 39.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.9137931034482759,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9192546583850932,
+      "eval_LOCATION_precision": 0.9137931034482759,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.9090909090909092,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9024390243902439,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.9006211180124224,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03947949409484863,
-      "eval_overall_accuracy": 0.9876475432335987,
-      "eval_overall_f1": 0.9427860696517413,
-      "eval_overall_precision": 0.9334975369458128,
-      "eval_overall_recall": 0.9522613065326633,
-      "eval_runtime": 0.6271,
-      "eval_samples_per_second": 271.101,
-      "eval_steps_per_second": 4.784,
+      "eval_loss": 0.036575764417648315,
+      "eval_overall_accuracy": 0.9884710403513588,
+      "eval_overall_f1": 0.9349999999999999,
+      "eval_overall_precision": 0.9303482587064676,
+      "eval_overall_recall": 0.9396984924623115,
+      "eval_runtime": 0.2834,
+      "eval_samples_per_second": 599.755,
+      "eval_steps_per_second": 10.584,
       "step": 3744
     },
     {
       "epoch": 40.0,
-      "grad_norm": 1.1784359216690063,
+      "grad_norm": 0.8054441809654236,
       "learning_rate": 3e-05,
-      "loss": 0.0484,
+      "loss": 0.0472,
       "step": 3840
     },
     {
       "epoch": 40.0,
-      "eval_LOCATION_f1": 0.9184549356223175,
+      "eval_LOCATION_f1": 0.9137931034482759,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9145299145299145,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9015873015873016,
+      "eval_LOCATION_precision": 0.9137931034482759,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8987341772151899,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9044585987261147,
+      "eval_ORGANIZATION_precision": 0.8987341772151899,
       "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03872258961200714,
-      "eval_overall_accuracy": 0.9862750480373319,
-      "eval_overall_f1": 0.9335006273525721,
-      "eval_overall_precision": 0.9323308270676691,
-      "eval_overall_recall": 0.9346733668341709,
-      "eval_runtime": 0.6311,
-      "eval_samples_per_second": 269.362,
-      "eval_steps_per_second": 4.753,
+      "eval_loss": 0.03671231493353844,
+      "eval_overall_accuracy": 0.9868240461158386,
+      "eval_overall_f1": 0.9309912170639899,
+      "eval_overall_precision": 0.9298245614035088,
+      "eval_overall_recall": 0.9321608040201005,
+      "eval_runtime": 0.2821,
+      "eval_samples_per_second": 602.601,
+      "eval_steps_per_second": 10.634,
       "step": 3840
     },
     {
       "epoch": 41.0,
-      "grad_norm": 0.21162720024585724,
+      "grad_norm": 0.3345367908477783,
       "learning_rate": 2.95e-05,
-      "loss": 0.0482,
+      "loss": 0.0486,
       "step": 3936
     },
     {
       "epoch": 41.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9191489361702128,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.907563025210084,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9249999999999999,
+      "eval_ORGANIZATION_f1": 0.91875,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9135802469135802,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.9074074074074074,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.038246579468250275,
-      "eval_overall_accuracy": 0.9876475432335987,
-      "eval_overall_f1": 0.9463171036204745,
-      "eval_overall_precision": 0.9404466501240695,
-      "eval_overall_recall": 0.9522613065326633,
-      "eval_runtime": 0.6407,
-      "eval_samples_per_second": 265.354,
-      "eval_steps_per_second": 4.683,
+      "eval_loss": 0.038836024701595306,
+      "eval_overall_accuracy": 0.9881965413121054,
+      "eval_overall_f1": 0.9402985074626866,
+      "eval_overall_precision": 0.9310344827586207,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2815,
+      "eval_samples_per_second": 603.989,
+      "eval_steps_per_second": 10.659,
       "step": 3936
     },
     {
       "epoch": 42.0,
-      "grad_norm": 0.4566030502319336,
+      "grad_norm": 0.16970324516296387,
       "learning_rate": 2.9e-05,
-      "loss": 0.0466,
+      "loss": 0.047,
       "step": 4032
     },
     {
       "epoch": 42.0,
-      "eval_LOCATION_f1": 0.923076923076923,
+      "eval_LOCATION_f1": 0.9145299145299146,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9152542372881356,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.9067796610169492,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.9073482428115016,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9161290322580645,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03999612480401993,
-      "eval_overall_accuracy": 0.987098545155092,
-      "eval_overall_f1": 0.9425,
-      "eval_overall_precision": 0.9378109452736318,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6254,
-      "eval_samples_per_second": 271.821,
-      "eval_steps_per_second": 4.797,
+      "eval_loss": 0.037543974816799164,
+      "eval_overall_accuracy": 0.9873730441943454,
+      "eval_overall_f1": 0.9346733668341709,
+      "eval_overall_precision": 0.9346733668341709,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.28,
+      "eval_samples_per_second": 607.074,
+      "eval_steps_per_second": 10.713,
       "step": 4032
     },
     {
       "epoch": 43.0,
-      "grad_norm": 1.2713156938552856,
+      "grad_norm": 0.6317854523658752,
       "learning_rate": 2.8499999999999998e-05,
-      "loss": 0.0461,
+      "loss": 0.0481,
       "step": 4128
     },
     {
       "epoch": 43.0,
-      "eval_LOCATION_f1": 0.9137931034482759,
+      "eval_LOCATION_f1": 0.9059829059829059,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9137931034482759,
+      "eval_LOCATION_precision": 0.8983050847457628,
       "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.9022082018927445,
+      "eval_ORGANIZATION_f1": 0.9050632911392406,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.89937106918239,
+      "eval_ORGANIZATION_precision": 0.9050632911392406,
       "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03929824382066727,
-      "eval_overall_accuracy": 0.9868240461158386,
-      "eval_overall_f1": 0.9323308270676693,
-      "eval_overall_precision": 0.93,
+      "eval_loss": 0.03800812363624573,
+      "eval_overall_accuracy": 0.9860005489980785,
+      "eval_overall_f1": 0.9311639549436797,
+      "eval_overall_precision": 0.9276807980049875,
       "eval_overall_recall": 0.9346733668341709,
-      "eval_runtime": 0.6205,
-      "eval_samples_per_second": 273.958,
-      "eval_steps_per_second": 4.835,
+      "eval_runtime": 0.2816,
+      "eval_samples_per_second": 603.734,
+      "eval_steps_per_second": 10.654,
       "step": 4128
     },
     {
       "epoch": 44.0,
-      "grad_norm": 0.5311025977134705,
+      "grad_norm": 0.9596404433250427,
       "learning_rate": 2.8000000000000003e-05,
-      "loss": 0.0462,
+      "loss": 0.0468,
       "step": 4224
     },
     {
       "epoch": 44.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.9230769230769231,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9171974522292994,
+      "eval_ORGANIZATION_f1": 0.9119496855345911,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9230769230769231,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
-      "eval_PERSON_f1": 0.976,
+      "eval_ORGANIZATION_precision": 0.90625,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9682539682539683,
-      "eval_PERSON_recall": 0.9838709677419355,
-      "eval_loss": 0.03938368707895279,
-      "eval_overall_accuracy": 0.9865495470765853,
-      "eval_overall_f1": 0.9396984924623115,
-      "eval_overall_precision": 0.9396984924623115,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.627,
-      "eval_samples_per_second": 271.145,
-      "eval_steps_per_second": 4.785,
+      "eval_PERSON_precision": 0.984,
+      "eval_PERSON_recall": 0.9919354838709677,
+      "eval_loss": 0.03910643607378006,
+      "eval_overall_accuracy": 0.9876475432335987,
+      "eval_overall_f1": 0.94,
+      "eval_overall_precision": 0.9353233830845771,
+      "eval_overall_recall": 0.9447236180904522,
+      "eval_runtime": 0.2799,
+      "eval_samples_per_second": 607.283,
+      "eval_steps_per_second": 10.717,
       "step": 4224
     },
     {
       "epoch": 45.0,
-      "grad_norm": 1.5404696464538574,
+      "grad_norm": 1.3304390907287598,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.0447,
+      "loss": 0.0473,
       "step": 4320
     },
     {
       "epoch": 45.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9106382978723405,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
+      "eval_LOCATION_precision": 0.8991596638655462,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_ORGANIZATION_f1": 0.9015873015873016,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9044585987261147,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03845032677054405,
-      "eval_overall_accuracy": 0.9873730441943454,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6186,
-      "eval_samples_per_second": 274.813,
-      "eval_steps_per_second": 4.85,
+      "eval_loss": 0.03656316548585892,
+      "eval_overall_accuracy": 0.9868240461158386,
+      "eval_overall_f1": 0.9311639549436797,
+      "eval_overall_precision": 0.9276807980049875,
+      "eval_overall_recall": 0.9346733668341709,
+      "eval_runtime": 0.2813,
+      "eval_samples_per_second": 604.337,
+      "eval_steps_per_second": 10.665,
       "step": 4320
     },
     {
       "epoch": 46.0,
-      "grad_norm": 0.4397522807121277,
+      "grad_norm": 0.8157849907875061,
       "learning_rate": 2.7000000000000002e-05,
-      "loss": 0.0442,
+      "loss": 0.0441,
       "step": 4416
     },
     {
       "epoch": 46.0,
-      "eval_LOCATION_f1": 0.935064935064935,
+      "eval_LOCATION_f1": 0.9152542372881356,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_precision": 0.9,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9182389937106918,
+      "eval_ORGANIZATION_f1": 0.9090909090909092,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9125,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9006211180124224,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03776326775550842,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9448621553884713,
-      "eval_overall_precision": 0.9425,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6325,
-      "eval_samples_per_second": 268.781,
-      "eval_steps_per_second": 4.743,
+      "eval_loss": 0.03721893951296806,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.9353233830845771,
+      "eval_overall_precision": 0.9261083743842364,
+      "eval_overall_recall": 0.9447236180904522,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 603.43,
+      "eval_steps_per_second": 10.649,
       "step": 4416
     },
     {
       "epoch": 47.0,
-      "grad_norm": 1.0225719213485718,
+      "grad_norm": 1.0967856645584106,
       "learning_rate": 2.6500000000000004e-05,
-      "loss": 0.0456,
+      "loss": 0.0441,
       "step": 4512
     },
     {
       "epoch": 47.0,
-      "eval_LOCATION_f1": 0.9304347826086956,
+      "eval_LOCATION_f1": 0.9224137931034483,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9385964912280702,
+      "eval_LOCATION_precision": 0.9224137931034483,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9316770186335404,
+      "eval_ORGANIZATION_f1": 0.9153605015673981,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9146341463414634,
-      "eval_ORGANIZATION_recall": 0.9493670886075949,
+      "eval_ORGANIZATION_precision": 0.906832298136646,
+      "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037870533764362335,
-      "eval_overall_accuracy": 0.9892945374691189,
-      "eval_overall_f1": 0.9488139825218478,
-      "eval_overall_precision": 0.9429280397022333,
-      "eval_overall_recall": 0.9547738693467337,
-      "eval_runtime": 0.6338,
-      "eval_samples_per_second": 268.242,
-      "eval_steps_per_second": 4.734,
+      "eval_loss": 0.037474945187568665,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.94,
+      "eval_overall_precision": 0.9353233830845771,
+      "eval_overall_recall": 0.9447236180904522,
+      "eval_runtime": 0.2834,
+      "eval_samples_per_second": 599.905,
+      "eval_steps_per_second": 10.587,
       "step": 4512
     },
     {
       "epoch": 48.0,
-      "grad_norm": 0.7907796502113342,
+      "grad_norm": 0.8679947853088379,
       "learning_rate": 2.6000000000000002e-05,
-      "loss": 0.0444,
+      "loss": 0.0416,
       "step": 4608
     },
     {
       "epoch": 48.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9316239316239315,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9182389937106918,
+      "eval_LOCATION_precision": 0.923728813559322,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9333333333333332,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9125,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9363057324840764,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03684180602431297,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9436795994993743,
-      "eval_overall_precision": 0.940149625935162,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6242,
-      "eval_samples_per_second": 272.342,
-      "eval_steps_per_second": 4.806,
+      "eval_loss": 0.03589639067649841,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9498746867167919,
+      "eval_overall_precision": 0.9475,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2828,
+      "eval_samples_per_second": 601.064,
+      "eval_steps_per_second": 10.607,
       "step": 4608
     },
     {
       "epoch": 49.0,
-      "grad_norm": 1.0734015703201294,
+      "grad_norm": 0.5795214176177979,
       "learning_rate": 2.5500000000000003e-05,
-      "loss": 0.045,
+      "loss": 0.0446,
       "step": 4704
     },
     {
       "epoch": 49.0,
-      "eval_LOCATION_f1": 0.9316239316239315,
+      "eval_LOCATION_f1": 0.923076923076923,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.923728813559322,
-      "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.9152542372881356,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.8958990536277602,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.8930817610062893,
+      "eval_ORGANIZATION_recall": 0.8987341772151899,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03768336772918701,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9461827284105131,
-      "eval_overall_precision": 0.942643391521197,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.623,
-      "eval_samples_per_second": 272.878,
-      "eval_steps_per_second": 4.815,
+      "eval_loss": 0.03545619174838066,
+      "eval_overall_accuracy": 0.9876475432335987,
+      "eval_overall_f1": 0.9325,
+      "eval_overall_precision": 0.927860696517413,
+      "eval_overall_recall": 0.9371859296482412,
+      "eval_runtime": 0.2807,
+      "eval_samples_per_second": 605.543,
+      "eval_steps_per_second": 10.686,
       "step": 4704
     },
     {
       "epoch": 50.0,
-      "grad_norm": 1.0651848316192627,
+      "grad_norm": 1.6378672122955322,
       "learning_rate": 2.5e-05,
-      "loss": 0.0461,
+      "loss": 0.0425,
       "step": 4800
     },
     {
       "epoch": 50.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.9276595744680851,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9177215189873418,
+      "eval_LOCATION_precision": 0.9159663865546218,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9056603773584907,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9177215189873418,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9,
+      "eval_ORGANIZATION_recall": 0.9113924050632911,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03915992006659508,
-      "eval_overall_accuracy": 0.987098545155092,
-      "eval_overall_f1": 0.9410288582183186,
-      "eval_overall_precision": 0.9398496240601504,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6242,
-      "eval_samples_per_second": 272.362,
-      "eval_steps_per_second": 4.806,
+      "eval_loss": 0.036550432443618774,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.9376558603491272,
+      "eval_overall_precision": 0.9306930693069307,
+      "eval_overall_recall": 0.9447236180904522,
+      "eval_runtime": 0.2824,
+      "eval_samples_per_second": 601.912,
+      "eval_steps_per_second": 10.622,
       "step": 4800
     },
     {
       "epoch": 51.0,
-      "grad_norm": 0.6373057961463928,
+      "grad_norm": 0.6017678380012512,
       "learning_rate": 2.45e-05,
       "loss": 0.0422,
       "step": 4896
     },
     {
       "epoch": 51.0,
-      "eval_LOCATION_f1": 0.9356223175965666,
+      "eval_LOCATION_f1": 0.923076923076923,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9316239316239316,
-      "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9177215189873418,
+      "eval_LOCATION_precision": 0.9152542372881356,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.910828025477707,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9177215189873418,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9166666666666666,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03710971772670746,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9448621553884713,
-      "eval_overall_precision": 0.9425,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6344,
-      "eval_samples_per_second": 267.952,
-      "eval_steps_per_second": 4.729,
+      "eval_loss": 0.036360133439302444,
+      "eval_overall_accuracy": 0.987098545155092,
+      "eval_overall_f1": 0.9385194479297366,
+      "eval_overall_precision": 0.9373433583959899,
+      "eval_overall_recall": 0.9396984924623115,
+      "eval_runtime": 0.2813,
+      "eval_samples_per_second": 604.3,
+      "eval_steps_per_second": 10.664,
       "step": 4896
     },
     {
       "epoch": 52.0,
-      "grad_norm": 1.6656302213668823,
+      "grad_norm": 1.1220182180404663,
       "learning_rate": 2.4e-05,
-      "loss": 0.0419,
+      "loss": 0.0409,
       "step": 4992
     },
     {
       "epoch": 52.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.9184549356223175,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9153605015673981,
-      "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.906832298136646,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.9145299145299145,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.91875,
+      "eval_ORGANIZATION_number": 158,
+      "eval_ORGANIZATION_precision": 0.9074074074074074,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03852337226271629,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9413233458177278,
-      "eval_overall_precision": 0.9354838709677419,
+      "eval_loss": 0.03568459302186966,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9401496259351622,
+      "eval_overall_precision": 0.9331683168316832,
       "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6311,
-      "eval_samples_per_second": 269.377,
-      "eval_steps_per_second": 4.754,
+      "eval_runtime": 0.2809,
+      "eval_samples_per_second": 605.214,
+      "eval_steps_per_second": 10.68,
       "step": 4992
     },
     {
       "epoch": 53.0,
-      "grad_norm": 1.0709712505340576,
+      "grad_norm": 1.0612504482269287,
       "learning_rate": 2.35e-05,
-      "loss": 0.0403,
+      "loss": 0.0414,
       "step": 5088
     },
     {
       "epoch": 53.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9182389937106918,
+      "eval_LOCATION_precision": 0.9230769230769231,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9249999999999999,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9125,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9135802469135802,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03830442577600479,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9423558897243106,
-      "eval_overall_precision": 0.94,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6351,
-      "eval_samples_per_second": 267.692,
-      "eval_steps_per_second": 4.724,
+      "eval_loss": 0.03593689948320389,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9451371571072318,
+      "eval_overall_precision": 0.9381188118811881,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2864,
+      "eval_samples_per_second": 593.537,
+      "eval_steps_per_second": 10.474,
       "step": 5088
     },
     {
       "epoch": 54.0,
-      "grad_norm": 1.4942684173583984,
+      "grad_norm": 1.130049228668213,
       "learning_rate": 2.3000000000000003e-05,
-      "loss": 0.0392,
+      "loss": 0.0403,
       "step": 5184
     },
     {
       "epoch": 54.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9221183800623053,
+      "eval_LOCATION_precision": 0.9230769230769231,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9130434782608695,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9079754601226994,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.8963414634146342,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037072986364364624,
-      "eval_overall_accuracy": 0.9898435355476256,
-      "eval_overall_f1": 0.9426433915211971,
-      "eval_overall_precision": 0.9356435643564357,
+      "eval_loss": 0.03527674823999405,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9402985074626866,
+      "eval_overall_precision": 0.9310344827586207,
       "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.622,
-      "eval_samples_per_second": 273.303,
-      "eval_steps_per_second": 4.823,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.883,
+      "eval_steps_per_second": 10.692,
       "step": 5184
     },
     {
       "epoch": 55.0,
-      "grad_norm": 2.575190544128418,
+      "grad_norm": 1.4285669326782227,
       "learning_rate": 2.25e-05,
-      "loss": 0.0419,
+      "loss": 0.0393,
       "step": 5280
     },
     {
       "epoch": 55.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9184549356223175,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.9145299145299145,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.9249999999999999,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9135802469135802,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03656064346432686,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9448621553884713,
-      "eval_overall_precision": 0.9425,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6256,
-      "eval_samples_per_second": 271.758,
-      "eval_steps_per_second": 4.796,
+      "eval_loss": 0.03516939654946327,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9426433915211971,
+      "eval_overall_precision": 0.9356435643564357,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 603.519,
+      "eval_steps_per_second": 10.65,
       "step": 5280
     },
     {
       "epoch": 56.0,
-      "grad_norm": 0.9610713720321655,
+      "grad_norm": 0.668100893497467,
       "learning_rate": 2.2000000000000003e-05,
-      "loss": 0.0391,
+      "loss": 0.0405,
       "step": 5376
     },
     {
       "epoch": 56.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9316239316239315,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.923728813559322,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9430379746835443,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9430379746835443,
+      "eval_ORGANIZATION_recall": 0.9430379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03741344437003136,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9448621553884713,
-      "eval_overall_precision": 0.9425,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6345,
-      "eval_samples_per_second": 267.915,
-      "eval_steps_per_second": 4.728,
+      "eval_loss": 0.03589174896478653,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9536921151439299,
+      "eval_overall_precision": 0.9501246882793017,
+      "eval_overall_recall": 0.957286432160804,
+      "eval_runtime": 0.2822,
+      "eval_samples_per_second": 602.42,
+      "eval_steps_per_second": 10.631,
       "step": 5376
     },
     {
       "epoch": 57.0,
-      "grad_norm": 0.37668493390083313,
+      "grad_norm": 0.7897509336471558,
       "learning_rate": 2.15e-05,
-      "loss": 0.0381,
+      "loss": 0.0404,
       "step": 5472
     },
     {
       "epoch": 57.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9276595744680851,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9278996865203761,
+      "eval_LOCATION_precision": 0.9159663865546218,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9400630914826499,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9192546583850931,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.9371069182389937,
+      "eval_ORGANIZATION_recall": 0.9430379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03578212484717369,
-      "eval_overall_accuracy": 0.9898435355476256,
-      "eval_overall_f1": 0.9475,
-      "eval_overall_precision": 0.9427860696517413,
-      "eval_overall_recall": 0.9522613065326633,
-      "eval_runtime": 0.6253,
-      "eval_samples_per_second": 271.877,
-      "eval_steps_per_second": 4.798,
+      "eval_loss": 0.037044160068035126,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.951310861423221,
+      "eval_overall_precision": 0.9454094292803971,
+      "eval_overall_recall": 0.957286432160804,
+      "eval_runtime": 0.2798,
+      "eval_samples_per_second": 607.511,
+      "eval_steps_per_second": 10.721,
       "step": 5472
     },
     {
       "epoch": 58.0,
-      "grad_norm": 0.46238428354263306,
+      "grad_norm": 0.9230952262878418,
       "learning_rate": 2.1e-05,
-      "loss": 0.0385,
+      "loss": 0.0398,
       "step": 5568
     },
     {
       "epoch": 58.0,
-      "eval_LOCATION_f1": 0.9396551724137931,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9396551724137931,
+      "eval_LOCATION_precision": 0.9316239316239316,
       "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9299363057324842,
+      "eval_ORGANIZATION_f1": 0.9337539432176657,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9358974358974359,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9308176100628931,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037841230630874634,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.950943396226415,
-      "eval_overall_precision": 0.9521410579345088,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.682,
-      "eval_samples_per_second": 249.282,
-      "eval_steps_per_second": 4.399,
+      "eval_loss": 0.035523511469364166,
+      "eval_overall_accuracy": 0.9903925336261323,
+      "eval_overall_f1": 0.951188986232791,
+      "eval_overall_precision": 0.9476309226932669,
+      "eval_overall_recall": 0.9547738693467337,
+      "eval_runtime": 0.282,
+      "eval_samples_per_second": 602.9,
+      "eval_steps_per_second": 10.639,
       "step": 5568
     },
     {
       "epoch": 59.0,
-      "grad_norm": 1.334747314453125,
+      "grad_norm": 0.6653354167938232,
       "learning_rate": 2.05e-05,
-      "loss": 0.036,
+      "loss": 0.0382,
       "step": 5664
     },
     {
       "epoch": 59.0,
-      "eval_LOCATION_f1": 0.9184549356223175,
+      "eval_LOCATION_f1": 0.9396551724137931,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9145299145299145,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.9396551724137931,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9490445859872612,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9551282051282052,
+      "eval_ORGANIZATION_recall": 0.9430379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03871507570147514,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9411764705882354,
-      "eval_overall_precision": 0.9376558603491272,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.626,
-      "eval_samples_per_second": 271.565,
-      "eval_steps_per_second": 4.792,
+      "eval_loss": 0.03545346483588219,
+      "eval_overall_accuracy": 0.9903925336261323,
+      "eval_overall_f1": 0.9584905660377357,
+      "eval_overall_precision": 0.9596977329974811,
+      "eval_overall_recall": 0.957286432160804,
+      "eval_runtime": 0.285,
+      "eval_samples_per_second": 596.488,
+      "eval_steps_per_second": 10.526,
       "step": 5664
     },
     {
       "epoch": 60.0,
-      "grad_norm": 1.4887285232543945,
+      "grad_norm": 1.0160192251205444,
       "learning_rate": 2e-05,
-      "loss": 0.0401,
+      "loss": 0.0396,
       "step": 5760
     },
     {
       "epoch": 60.0,
-      "eval_LOCATION_f1": 0.9356223175965666,
+      "eval_LOCATION_f1": 0.9276595744680851,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_precision": 0.9159663865546218,
       "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_ORGANIZATION_f1": 0.9182389937106918,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
+      "eval_ORGANIZATION_precision": 0.9125,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03643361106514931,
-      "eval_overall_accuracy": 0.9890200384298655,
-      "eval_overall_f1": 0.9461827284105131,
-      "eval_overall_precision": 0.942643391521197,
+      "eval_loss": 0.03439299017190933,
+      "eval_overall_accuracy": 0.9892945374691189,
+      "eval_overall_f1": 0.9426433915211971,
+      "eval_overall_precision": 0.9356435643564357,
       "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6277,
-      "eval_samples_per_second": 270.809,
-      "eval_steps_per_second": 4.779,
+      "eval_runtime": 0.2803,
+      "eval_samples_per_second": 606.413,
+      "eval_steps_per_second": 10.701,
       "step": 5760
     },
     {
       "epoch": 61.0,
-      "grad_norm": 0.551049530506134,
+      "grad_norm": 0.5373579263687134,
       "learning_rate": 1.9500000000000003e-05,
-      "loss": 0.0357,
+      "loss": 0.0362,
       "step": 5856
     },
     {
       "epoch": 61.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.9230769230769231,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_ORGANIZATION_f1": 0.9137380191693291,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9225806451612903,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03617864474654198,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9447236180904522,
-      "eval_overall_precision": 0.9447236180904522,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6313,
-      "eval_samples_per_second": 269.281,
-      "eval_steps_per_second": 4.752,
+      "eval_loss": 0.03558684140443802,
+      "eval_overall_accuracy": 0.9879220422728521,
+      "eval_overall_f1": 0.9408805031446541,
+      "eval_overall_precision": 0.9420654911838791,
+      "eval_overall_recall": 0.9396984924623115,
+      "eval_runtime": 0.2814,
+      "eval_samples_per_second": 604.106,
+      "eval_steps_per_second": 10.661,
       "step": 5856
     },
     {
       "epoch": 62.0,
-      "grad_norm": 0.6859171986579895,
+      "grad_norm": 1.3705915212631226,
       "learning_rate": 1.9e-05,
-      "loss": 0.0342,
+      "loss": 0.037,
       "step": 5952
     },
     {
@@ -1912,192 +1912,192 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.923728813559322,
       "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_ORGANIZATION_f1": 0.910828025477707,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
-      "eval_PERSON_f1": 0.9840000000000001,
+      "eval_ORGANIZATION_precision": 0.9166666666666666,
+      "eval_ORGANIZATION_recall": 0.9050632911392406,
+      "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
-      "eval_PERSON_precision": 0.9761904761904762,
+      "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03858007490634918,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9450000000000001,
-      "eval_overall_precision": 0.9402985074626866,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.63,
-      "eval_samples_per_second": 269.861,
-      "eval_steps_per_second": 4.762,
+      "eval_loss": 0.03602343425154686,
+      "eval_overall_accuracy": 0.9881965413121054,
+      "eval_overall_f1": 0.9410288582183186,
+      "eval_overall_precision": 0.9398496240601504,
+      "eval_overall_recall": 0.9422110552763819,
+      "eval_runtime": 0.2805,
+      "eval_samples_per_second": 606.05,
+      "eval_steps_per_second": 10.695,
       "step": 5952
     },
     {
       "epoch": 63.0,
-      "grad_norm": 0.6219615340232849,
+      "grad_norm": 0.5670416355133057,
       "learning_rate": 1.85e-05,
-      "loss": 0.0363,
+      "loss": 0.0386,
       "step": 6048
     },
     {
       "epoch": 63.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.9310344827586207,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
+      "eval_LOCATION_precision": 0.9310344827586207,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.930379746835443,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.930379746835443,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03788018971681595,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9473684210526316,
-      "eval_overall_precision": 0.945,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6257,
-      "eval_samples_per_second": 271.683,
-      "eval_steps_per_second": 4.794,
+      "eval_loss": 0.036378536373376846,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9510664993726474,
+      "eval_overall_precision": 0.949874686716792,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2812,
+      "eval_samples_per_second": 604.53,
+      "eval_steps_per_second": 10.668,
       "step": 6048
     },
     {
       "epoch": 64.0,
-      "grad_norm": 1.1038762331008911,
+      "grad_norm": 1.2524464130401611,
       "learning_rate": 1.8e-05,
-      "loss": 0.0355,
+      "loss": 0.0365,
       "step": 6144
     },
     {
       "epoch": 64.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.923076923076923,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.9152542372881356,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9142857142857144,
+      "eval_ORGANIZATION_f1": 0.9260450160771704,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9171974522292994,
+      "eval_ORGANIZATION_precision": 0.9411764705882353,
       "eval_ORGANIZATION_recall": 0.9113924050632911,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.036534231156110764,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9422110552763819,
-      "eval_overall_precision": 0.9422110552763819,
+      "eval_loss": 0.036032553762197495,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.9445843828715367,
+      "eval_overall_precision": 0.946969696969697,
       "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6276,
-      "eval_samples_per_second": 270.88,
-      "eval_steps_per_second": 4.78,
+      "eval_runtime": 0.282,
+      "eval_samples_per_second": 602.84,
+      "eval_steps_per_second": 10.638,
       "step": 6144
     },
     {
       "epoch": 65.0,
-      "grad_norm": 0.6325670480728149,
+      "grad_norm": 0.5007239580154419,
       "learning_rate": 1.75e-05,
-      "loss": 0.0353,
+      "loss": 0.0347,
       "step": 6240
     },
     {
       "epoch": 65.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9316239316239315,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.923728813559322,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9294871794871796,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9415584415584416,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03730296343564987,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9447236180904522,
-      "eval_overall_precision": 0.9447236180904522,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6323,
-      "eval_samples_per_second": 268.873,
-      "eval_steps_per_second": 4.745,
+      "eval_loss": 0.035410117357969284,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.9484276729559749,
+      "eval_overall_precision": 0.9496221662468514,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2862,
+      "eval_samples_per_second": 594.083,
+      "eval_steps_per_second": 10.484,
       "step": 6240
     },
     {
       "epoch": 66.0,
-      "grad_norm": 0.7639645934104919,
+      "grad_norm": 0.7251582741737366,
       "learning_rate": 1.7000000000000003e-05,
-      "loss": 0.0343,
+      "loss": 0.0393,
       "step": 6336
     },
     {
       "epoch": 66.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9396551724137931,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9308176100628932,
+      "eval_LOCATION_precision": 0.9396551724137931,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.926517571884984,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.925,
-      "eval_ORGANIZATION_recall": 0.9367088607594937,
+      "eval_ORGANIZATION_precision": 0.9354838709677419,
+      "eval_ORGANIZATION_recall": 0.9177215189873418,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03759036585688591,
+      "eval_loss": 0.036589812487363815,
       "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9473684210526316,
-      "eval_overall_precision": 0.945,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6238,
-      "eval_samples_per_second": 272.526,
-      "eval_steps_per_second": 4.809,
+      "eval_overall_f1": 0.9496221662468514,
+      "eval_overall_precision": 0.952020202020202,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2814,
+      "eval_samples_per_second": 604.074,
+      "eval_steps_per_second": 10.66,
       "step": 6336
     },
     {
       "epoch": 67.0,
-      "grad_norm": 1.125841736793518,
+      "grad_norm": 1.2316526174545288,
       "learning_rate": 1.65e-05,
-      "loss": 0.0347,
+      "loss": 0.0359,
       "step": 6432
     },
     {
       "epoch": 67.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9240506329113924,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
+      "eval_ORGANIZATION_precision": 0.9240506329113924,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03600994125008583,
-      "eval_overall_accuracy": 0.9890200384298655,
-      "eval_overall_f1": 0.9436795994993743,
-      "eval_overall_precision": 0.940149625935162,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6257,
-      "eval_samples_per_second": 271.681,
-      "eval_steps_per_second": 4.794,
+      "eval_loss": 0.034778717905282974,
+      "eval_overall_accuracy": 0.9892945374691189,
+      "eval_overall_f1": 0.9473684210526316,
+      "eval_overall_precision": 0.945,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2795,
+      "eval_samples_per_second": 608.215,
+      "eval_steps_per_second": 10.733,
       "step": 6432
     },
     {
       "epoch": 68.0,
-      "grad_norm": 1.1081820726394653,
+      "grad_norm": 0.5684135556221008,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.0339,
+      "loss": 0.0331,
       "step": 6528
     },
     {
       "epoch": 68.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
       "eval_ORGANIZATION_f1": 0.9177215189873418,
       "eval_ORGANIZATION_number": 158,
       "eval_ORGANIZATION_precision": 0.9177215189873418,
@@ -2106,83 +2106,83 @@
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03698847442865372,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9435382685069008,
-      "eval_overall_precision": 0.9423558897243107,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6327,
-      "eval_samples_per_second": 268.707,
-      "eval_steps_per_second": 4.742,
+      "eval_loss": 0.03474760055541992,
+      "eval_overall_accuracy": 0.9890200384298655,
+      "eval_overall_f1": 0.9448621553884713,
+      "eval_overall_precision": 0.9425,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2799,
+      "eval_samples_per_second": 607.468,
+      "eval_steps_per_second": 10.72,
       "step": 6528
     },
     {
       "epoch": 69.0,
-      "grad_norm": 0.5234565138816833,
+      "grad_norm": 0.7760136127471924,
       "learning_rate": 1.55e-05,
-      "loss": 0.0357,
+      "loss": 0.0344,
       "step": 6624
     },
     {
       "epoch": 69.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9142857142857144,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9333333333333332,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9171974522292994,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.9363057324840764,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03598235175013542,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9408805031446541,
-      "eval_overall_precision": 0.9420654911838791,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6389,
-      "eval_samples_per_second": 266.081,
-      "eval_steps_per_second": 4.696,
+      "eval_loss": 0.03407077118754387,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.950943396226415,
+      "eval_overall_precision": 0.9521410579345088,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2833,
+      "eval_samples_per_second": 600.147,
+      "eval_steps_per_second": 10.591,
       "step": 6624
     },
     {
       "epoch": 70.0,
-      "grad_norm": 1.084986686706543,
+      "grad_norm": 1.5560885667800903,
       "learning_rate": 1.5e-05,
-      "loss": 0.0339,
+      "loss": 0.0349,
       "step": 6720
     },
     {
       "epoch": 70.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9396551724137931,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.9396551724137931,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.035142358392477036,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9447236180904522,
-      "eval_overall_precision": 0.9447236180904522,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.628,
-      "eval_samples_per_second": 270.695,
-      "eval_steps_per_second": 4.777,
+      "eval_loss": 0.034501850605010986,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9547738693467337,
+      "eval_overall_precision": 0.9547738693467337,
+      "eval_overall_recall": 0.9547738693467337,
+      "eval_runtime": 0.28,
+      "eval_samples_per_second": 607.101,
+      "eval_steps_per_second": 10.714,
       "step": 6720
     },
     {
       "epoch": 71.0,
-      "grad_norm": 1.4800338745117188,
+      "grad_norm": 0.7581639885902405,
       "learning_rate": 1.45e-05,
-      "loss": 0.0354,
+      "loss": 0.0349,
       "step": 6816
     },
     {
@@ -2191,29 +2191,29 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.9310344827586207,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9299363057324842,
+      "eval_ORGANIZATION_f1": 0.926984126984127,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9358974358974359,
+      "eval_ORGANIZATION_precision": 0.9299363057324841,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03501693159341812,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9484276729559749,
-      "eval_overall_precision": 0.9496221662468514,
+      "eval_loss": 0.03540712222456932,
+      "eval_overall_accuracy": 0.9884710403513588,
+      "eval_overall_f1": 0.9472361809045227,
+      "eval_overall_precision": 0.9472361809045227,
       "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6364,
-      "eval_samples_per_second": 267.144,
-      "eval_steps_per_second": 4.714,
+      "eval_runtime": 0.2795,
+      "eval_samples_per_second": 608.239,
+      "eval_steps_per_second": 10.734,
       "step": 6816
     },
     {
       "epoch": 72.0,
-      "grad_norm": 0.46930524706840515,
+      "grad_norm": 0.30566951632499695,
       "learning_rate": 1.4000000000000001e-05,
-      "loss": 0.0345,
+      "loss": 0.0342,
       "step": 6912
     },
     {
@@ -2222,91 +2222,91 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.923728813559322,
       "eval_LOCATION_recall": 0.9396551724137931,
-      "eval_ORGANIZATION_f1": 0.9211356466876972,
+      "eval_ORGANIZATION_f1": 0.926984126984127,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9182389937106918,
+      "eval_ORGANIZATION_precision": 0.9299363057324841,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0342424102127552,
-      "eval_overall_accuracy": 0.9895690365083722,
-      "eval_overall_f1": 0.9450000000000001,
-      "eval_overall_precision": 0.9402985074626866,
+      "eval_loss": 0.034342918545007706,
+      "eval_overall_accuracy": 0.9887455393906122,
+      "eval_overall_f1": 0.9473684210526316,
+      "eval_overall_precision": 0.945,
       "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.626,
-      "eval_samples_per_second": 271.587,
-      "eval_steps_per_second": 4.793,
+      "eval_runtime": 0.2798,
+      "eval_samples_per_second": 607.48,
+      "eval_steps_per_second": 10.72,
       "step": 6912
     },
     {
       "epoch": 73.0,
-      "grad_norm": 1.7347519397735596,
+      "grad_norm": 1.0282853841781616,
       "learning_rate": 1.3500000000000001e-05,
-      "loss": 0.0346,
+      "loss": 0.0333,
       "step": 7008
     },
     {
       "epoch": 73.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9363057324840764,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9240506329113924,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9423076923076923,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9240506329113924,
+      "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03550710529088974,
+      "eval_loss": 0.03536517545580864,
       "eval_overall_accuracy": 0.9890200384298655,
-      "eval_overall_f1": 0.9496221662468514,
-      "eval_overall_precision": 0.952020202020202,
+      "eval_overall_f1": 0.9472361809045227,
+      "eval_overall_precision": 0.9472361809045227,
       "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6305,
-      "eval_samples_per_second": 269.612,
-      "eval_steps_per_second": 4.758,
+      "eval_runtime": 0.2817,
+      "eval_samples_per_second": 603.375,
+      "eval_steps_per_second": 10.648,
       "step": 7008
     },
     {
       "epoch": 74.0,
-      "grad_norm": 0.18666690587997437,
+      "grad_norm": 3.3504393100738525,
       "learning_rate": 1.3000000000000001e-05,
-      "loss": 0.0322,
+      "loss": 0.0332,
       "step": 7104
     },
     {
       "epoch": 74.0,
-      "eval_LOCATION_f1": 0.935064935064935,
+      "eval_LOCATION_f1": 0.927038626609442,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_precision": 0.9230769230769231,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.939297124600639,
+      "eval_ORGANIZATION_f1": 0.9240506329113924,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9483870967741935,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9240506329113924,
+      "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03586093708872795,
-      "eval_overall_accuracy": 0.9895690365083722,
-      "eval_overall_f1": 0.953341740226986,
-      "eval_overall_precision": 0.9569620253164557,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6297,
-      "eval_samples_per_second": 269.97,
-      "eval_steps_per_second": 4.764,
+      "eval_loss": 0.0346221849322319,
+      "eval_overall_accuracy": 0.9892945374691189,
+      "eval_overall_f1": 0.9448621553884713,
+      "eval_overall_precision": 0.9425,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2789,
+      "eval_samples_per_second": 609.459,
+      "eval_steps_per_second": 10.755,
       "step": 7104
     },
     {
       "epoch": 75.0,
-      "grad_norm": 1.8562440872192383,
+      "grad_norm": 0.7059272527694702,
       "learning_rate": 1.25e-05,
-      "loss": 0.034,
+      "loss": 0.0346,
       "step": 7200
     },
     {
@@ -2315,122 +2315,122 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.9310344827586207,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9329073482428115,
+      "eval_ORGANIZATION_f1": 0.9274447949526813,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9419354838709677,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9245283018867925,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03533979505300522,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9496221662468514,
-      "eval_overall_precision": 0.952020202020202,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6258,
-      "eval_samples_per_second": 271.655,
-      "eval_steps_per_second": 4.794,
+      "eval_loss": 0.0342298299074173,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9473684210526316,
+      "eval_overall_precision": 0.945,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2836,
+      "eval_samples_per_second": 599.44,
+      "eval_steps_per_second": 10.578,
       "step": 7200
     },
     {
       "epoch": 76.0,
-      "grad_norm": 0.6034942865371704,
+      "grad_norm": 1.0325454473495483,
       "learning_rate": 1.2e-05,
-      "loss": 0.0327,
+      "loss": 0.0334,
       "step": 7296
     },
     {
       "epoch": 76.0,
-      "eval_LOCATION_f1": 0.935064935064935,
+      "eval_LOCATION_f1": 0.9224137931034483,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9391304347826087,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9363057324840764,
+      "eval_LOCATION_precision": 0.9224137931034483,
+      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_ORGANIZATION_f1": 0.9308176100628932,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9423076923076923,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.925,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03656287491321564,
-      "eval_overall_accuracy": 0.9890200384298655,
-      "eval_overall_f1": 0.9521410579345088,
-      "eval_overall_precision": 0.9545454545454546,
+      "eval_loss": 0.0346398688852787,
+      "eval_overall_accuracy": 0.9903925336261323,
+      "eval_overall_f1": 0.9461827284105131,
+      "eval_overall_precision": 0.942643391521197,
       "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6291,
-      "eval_samples_per_second": 270.237,
-      "eval_steps_per_second": 4.769,
+      "eval_runtime": 0.2805,
+      "eval_samples_per_second": 605.977,
+      "eval_steps_per_second": 10.694,
       "step": 7296
     },
     {
       "epoch": 77.0,
-      "grad_norm": 0.7069413661956787,
+      "grad_norm": 0.531468391418457,
       "learning_rate": 1.1500000000000002e-05,
-      "loss": 0.032,
+      "loss": 0.034,
       "step": 7392
     },
     {
       "epoch": 77.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9396551724137931,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.9396551724137931,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.926984126984127,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9299363057324841,
       "eval_ORGANIZATION_recall": 0.9240506329113924,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03621837496757507,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9447236180904522,
-      "eval_overall_precision": 0.9447236180904522,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6257,
-      "eval_samples_per_second": 271.697,
-      "eval_steps_per_second": 4.795,
+      "eval_loss": 0.03495112434029579,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.949748743718593,
+      "eval_overall_precision": 0.949748743718593,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2841,
+      "eval_samples_per_second": 598.465,
+      "eval_steps_per_second": 10.561,
       "step": 7392
     },
     {
       "epoch": 78.0,
-      "grad_norm": 0.9935917854309082,
+      "grad_norm": 0.8816949129104614,
       "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.0312,
+      "loss": 0.0341,
       "step": 7488
     },
     {
       "epoch": 78.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9201277955271565,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9333333333333332,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9290322580645162,
-      "eval_ORGANIZATION_recall": 0.9113924050632911,
+      "eval_ORGANIZATION_precision": 0.9363057324840764,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03663404658436775,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9420654911838791,
-      "eval_overall_precision": 0.9444444444444444,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6304,
-      "eval_samples_per_second": 269.69,
-      "eval_steps_per_second": 4.759,
+      "eval_loss": 0.034044817090034485,
+      "eval_overall_accuracy": 0.9903925336261323,
+      "eval_overall_f1": 0.9510664993726474,
+      "eval_overall_precision": 0.949874686716792,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2793,
+      "eval_samples_per_second": 608.583,
+      "eval_steps_per_second": 10.74,
       "step": 7488
     },
     {
       "epoch": 79.0,
-      "grad_norm": 0.24092857539653778,
+      "grad_norm": 0.5149128437042236,
       "learning_rate": 1.05e-05,
-      "loss": 0.03,
+      "loss": 0.033,
       "step": 7584
     },
     {
@@ -2439,60 +2439,60 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.9304347826086956,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_ORGANIZATION_f1": 0.9308176100628932,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.925,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03647984191775322,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9447236180904522,
-      "eval_overall_precision": 0.9447236180904522,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6175,
-      "eval_samples_per_second": 275.317,
-      "eval_steps_per_second": 4.859,
+      "eval_loss": 0.034786492586135864,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9473684210526316,
+      "eval_overall_precision": 0.945,
+      "eval_overall_recall": 0.949748743718593,
+      "eval_runtime": 0.2819,
+      "eval_samples_per_second": 602.992,
+      "eval_steps_per_second": 10.641,
       "step": 7584
     },
     {
       "epoch": 80.0,
-      "grad_norm": 0.6495091915130615,
+      "grad_norm": 1.0311975479125977,
       "learning_rate": 1e-05,
-      "loss": 0.0317,
+      "loss": 0.0308,
       "step": 7680
     },
     {
       "epoch": 80.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.9137931034482759,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9363057324840764,
+      "eval_LOCATION_precision": 0.9137931034482759,
+      "eval_LOCATION_recall": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.9278996865203761,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9423076923076923,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9192546583850931,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037438035011291504,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.950943396226415,
-      "eval_overall_precision": 0.9521410579345088,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6252,
-      "eval_samples_per_second": 271.933,
-      "eval_steps_per_second": 4.799,
+      "eval_loss": 0.0336722694337368,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9425,
+      "eval_overall_precision": 0.9378109452736318,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2809,
+      "eval_samples_per_second": 605.269,
+      "eval_steps_per_second": 10.681,
       "step": 7680
     },
     {
       "epoch": 81.0,
-      "grad_norm": 0.7272472381591797,
+      "grad_norm": 0.3829832971096039,
       "learning_rate": 9.5e-06,
-      "loss": 0.0324,
+      "loss": 0.031,
       "step": 7776
     },
     {
@@ -2501,37 +2501,37 @@
       "eval_LOCATION_number": 116,
       "eval_LOCATION_precision": 0.9224137931034483,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9337539432176657,
+      "eval_ORGANIZATION_f1": 0.9278996865203761,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9308176100628931,
+      "eval_ORGANIZATION_precision": 0.9192546583850931,
       "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03790738806128502,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9473684210526316,
-      "eval_overall_precision": 0.945,
+      "eval_loss": 0.03411416336894035,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9450000000000001,
+      "eval_overall_precision": 0.9402985074626866,
       "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.7303,
-      "eval_samples_per_second": 232.793,
-      "eval_steps_per_second": 4.108,
+      "eval_runtime": 0.2823,
+      "eval_samples_per_second": 602.183,
+      "eval_steps_per_second": 10.627,
       "step": 7776
     },
     {
       "epoch": 82.0,
-      "grad_norm": 0.4394896626472473,
+      "grad_norm": 0.5665716528892517,
       "learning_rate": 9e-06,
-      "loss": 0.0309,
+      "loss": 0.0315,
       "step": 7872
     },
     {
       "epoch": 82.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.9316239316239315,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
+      "eval_LOCATION_precision": 0.923728813559322,
+      "eval_LOCATION_recall": 0.9396551724137931,
       "eval_ORGANIZATION_f1": 0.9333333333333332,
       "eval_ORGANIZATION_number": 158,
       "eval_ORGANIZATION_precision": 0.9363057324840764,
@@ -2540,582 +2540,582 @@
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037485770881175995,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9472361809045227,
-      "eval_overall_precision": 0.9472361809045227,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.63,
-      "eval_samples_per_second": 269.821,
-      "eval_steps_per_second": 4.762,
+      "eval_loss": 0.03403652459383011,
+      "eval_overall_accuracy": 0.9903925336261323,
+      "eval_overall_f1": 0.9498746867167919,
+      "eval_overall_precision": 0.9475,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2825,
+      "eval_samples_per_second": 601.815,
+      "eval_steps_per_second": 10.62,
       "step": 7872
     },
     {
       "epoch": 83.0,
-      "grad_norm": 1.6514469385147095,
+      "grad_norm": 0.9708644151687622,
       "learning_rate": 8.500000000000002e-06,
-      "loss": 0.0333,
+      "loss": 0.0321,
       "step": 7968
     },
     {
       "epoch": 83.0,
-      "eval_LOCATION_f1": 0.9177489177489179,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9217391304347826,
-      "eval_LOCATION_recall": 0.9137931034482759,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03762008622288704,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9420654911838791,
-      "eval_overall_precision": 0.9444444444444444,
-      "eval_overall_recall": 0.9396984924623115,
-      "eval_runtime": 0.6209,
-      "eval_samples_per_second": 273.793,
-      "eval_steps_per_second": 4.832,
+      "eval_loss": 0.03434957191348076,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2812,
+      "eval_samples_per_second": 604.564,
+      "eval_steps_per_second": 10.669,
       "step": 7968
     },
     {
       "epoch": 84.0,
-      "grad_norm": 1.4008151292800903,
+      "grad_norm": 1.0318374633789062,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.0331,
+      "loss": 0.0317,
       "step": 8064
     },
     {
       "epoch": 84.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
+      "eval_LOCATION_precision": 0.9391304347826087,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9333333333333332,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9363057324840764,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.037528783082962036,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9485570890840652,
-      "eval_overall_precision": 0.9473684210526315,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6239,
-      "eval_samples_per_second": 272.485,
-      "eval_steps_per_second": 4.809,
+      "eval_loss": 0.03397549316287041,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2802,
+      "eval_samples_per_second": 606.677,
+      "eval_steps_per_second": 10.706,
       "step": 8064
     },
     {
       "epoch": 85.0,
-      "grad_norm": 0.8886189460754395,
+      "grad_norm": 1.0523338317871094,
       "learning_rate": 7.5e-06,
-      "loss": 0.0313,
+      "loss": 0.0324,
       "step": 8160
     },
     {
       "epoch": 85.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9184549356223175,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
+      "eval_LOCATION_precision": 0.9145299145299145,
       "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.926984126984127,
+      "eval_ORGANIZATION_f1": 0.9245283018867925,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9299363057324841,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.91875,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03677176684141159,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9459119496855346,
-      "eval_overall_precision": 0.947103274559194,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6275,
-      "eval_samples_per_second": 270.908,
-      "eval_steps_per_second": 4.781,
+      "eval_loss": 0.033996351063251495,
+      "eval_overall_accuracy": 0.9892945374691189,
+      "eval_overall_f1": 0.9425,
+      "eval_overall_precision": 0.9378109452736318,
+      "eval_overall_recall": 0.9472361809045227,
+      "eval_runtime": 0.2808,
+      "eval_samples_per_second": 605.465,
+      "eval_steps_per_second": 10.685,
       "step": 8160
     },
     {
       "epoch": 86.0,
-      "grad_norm": 0.4647173285484314,
+      "grad_norm": 0.45407137274742126,
       "learning_rate": 7.000000000000001e-06,
-      "loss": 0.031,
+      "loss": 0.0317,
       "step": 8256
     },
     {
       "epoch": 86.0,
-      "eval_LOCATION_f1": 0.927038626609442,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9230769230769231,
-      "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9333333333333332,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9363057324840764,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9363057324840764,
+      "eval_ORGANIZATION_precision": 0.9423076923076923,
       "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0371999628841877,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9485570890840652,
-      "eval_overall_precision": 0.9473684210526315,
-      "eval_overall_recall": 0.949748743718593,
-      "eval_runtime": 0.6241,
-      "eval_samples_per_second": 272.404,
-      "eval_steps_per_second": 4.807,
+      "eval_loss": 0.03390591964125633,
+      "eval_overall_accuracy": 0.990118034586879,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.825,
+      "eval_steps_per_second": 10.691,
       "step": 8256
     },
     {
       "epoch": 87.0,
-      "grad_norm": 0.5002410411834717,
+      "grad_norm": 0.26913008093833923,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.0325,
+      "loss": 0.0308,
       "step": 8352
     },
     {
       "epoch": 87.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9333333333333332,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.9363057324840764,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9363057324840764,
+      "eval_ORGANIZATION_precision": 0.9423076923076923,
       "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03664541244506836,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9472361809045227,
-      "eval_overall_precision": 0.9472361809045227,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6476,
-      "eval_samples_per_second": 262.503,
-      "eval_steps_per_second": 4.632,
+      "eval_loss": 0.03465632349252701,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.28,
+      "eval_samples_per_second": 607.249,
+      "eval_steps_per_second": 10.716,
       "step": 8352
     },
     {
       "epoch": 88.0,
-      "grad_norm": 1.6027255058288574,
+      "grad_norm": 0.30417150259017944,
       "learning_rate": 6e-06,
-      "loss": 0.0309,
+      "loss": 0.0311,
       "step": 8448
     },
     {
       "epoch": 88.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9333333333333332,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9363057324840764,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0366690568625927,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9472361809045227,
-      "eval_overall_precision": 0.9472361809045227,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6246,
-      "eval_samples_per_second": 272.173,
-      "eval_steps_per_second": 4.803,
+      "eval_loss": 0.034424372017383575,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2814,
+      "eval_samples_per_second": 604.076,
+      "eval_steps_per_second": 10.66,
       "step": 8448
     },
     {
       "epoch": 89.0,
-      "grad_norm": 1.264343023300171,
+      "grad_norm": 0.7772185206413269,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.0298,
+      "loss": 0.0295,
       "step": 8544
     },
     {
       "epoch": 89.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03642358258366585,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9445843828715367,
-      "eval_overall_precision": 0.946969696969697,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6228,
-      "eval_samples_per_second": 272.971,
-      "eval_steps_per_second": 4.817,
+      "eval_loss": 0.034607380628585815,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2806,
+      "eval_samples_per_second": 605.778,
+      "eval_steps_per_second": 10.69,
       "step": 8544
     },
     {
       "epoch": 90.0,
-      "grad_norm": 1.2060728073120117,
+      "grad_norm": 0.6779617071151733,
       "learning_rate": 5e-06,
-      "loss": 0.03,
+      "loss": 0.0304,
       "step": 8640
     },
     {
       "epoch": 90.0,
-      "eval_LOCATION_f1": 0.9310344827586207,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9310344827586207,
+      "eval_LOCATION_precision": 0.9391304347826087,
       "eval_LOCATION_recall": 0.9310344827586207,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03646455332636833,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9459119496855346,
-      "eval_overall_precision": 0.947103274559194,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.623,
-      "eval_samples_per_second": 272.869,
-      "eval_steps_per_second": 4.815,
+      "eval_loss": 0.03426329791545868,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2819,
+      "eval_samples_per_second": 603.089,
+      "eval_steps_per_second": 10.643,
       "step": 8640
     },
     {
       "epoch": 91.0,
-      "grad_norm": 1.7125660181045532,
+      "grad_norm": 1.0395246744155884,
       "learning_rate": 4.5e-06,
-      "loss": 0.0317,
+      "loss": 0.0315,
       "step": 8736
     },
     {
       "epoch": 91.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03602577745914459,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6244,
-      "eval_samples_per_second": 272.252,
-      "eval_steps_per_second": 4.804,
+      "eval_loss": 0.03426933288574219,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2786,
+      "eval_samples_per_second": 610.112,
+      "eval_steps_per_second": 10.767,
       "step": 8736
     },
     {
       "epoch": 92.0,
-      "grad_norm": 2.947300434112549,
+      "grad_norm": 0.45442330837249756,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.0308,
+      "loss": 0.0314,
       "step": 8832
     },
     {
       "epoch": 92.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03611500561237335,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9445843828715367,
-      "eval_overall_precision": 0.946969696969697,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6314,
-      "eval_samples_per_second": 269.227,
-      "eval_steps_per_second": 4.751,
+      "eval_loss": 0.03421960771083832,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.282,
+      "eval_samples_per_second": 602.845,
+      "eval_steps_per_second": 10.638,
       "step": 8832
     },
     {
       "epoch": 93.0,
-      "grad_norm": 0.5695700645446777,
+      "grad_norm": 0.4602469801902771,
       "learning_rate": 3.5000000000000004e-06,
-      "loss": 0.0282,
+      "loss": 0.0322,
       "step": 8928
     },
     {
       "epoch": 93.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.035832908004522324,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9445843828715367,
-      "eval_overall_precision": 0.946969696969697,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6207,
-      "eval_samples_per_second": 273.9,
-      "eval_steps_per_second": 4.834,
+      "eval_loss": 0.03397959843277931,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2823,
+      "eval_samples_per_second": 602.296,
+      "eval_steps_per_second": 10.629,
       "step": 8928
     },
     {
       "epoch": 94.0,
-      "grad_norm": 1.4625869989395142,
+      "grad_norm": 0.46450167894363403,
       "learning_rate": 3e-06,
-      "loss": 0.03,
+      "loss": 0.0303,
       "step": 9024
     },
     {
       "epoch": 94.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9333333333333332,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9363057324840764,
-      "eval_ORGANIZATION_recall": 0.930379746835443,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03618810325860977,
-      "eval_overall_accuracy": 0.9887455393906122,
-      "eval_overall_f1": 0.9472361809045227,
-      "eval_overall_precision": 0.9472361809045227,
-      "eval_overall_recall": 0.9472361809045227,
-      "eval_runtime": 0.6205,
-      "eval_samples_per_second": 273.954,
-      "eval_steps_per_second": 4.834,
+      "eval_loss": 0.03426538035273552,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.281,
+      "eval_samples_per_second": 604.875,
+      "eval_steps_per_second": 10.674,
       "step": 9024
     },
     {
       "epoch": 95.0,
-      "grad_norm": 0.5287604331970215,
+      "grad_norm": 0.763189971446991,
       "learning_rate": 2.5e-06,
-      "loss": 0.0318,
+      "loss": 0.0316,
       "step": 9120
     },
     {
       "epoch": 95.0,
-      "eval_LOCATION_f1": 0.9224137931034483,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9224137931034483,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9240506329113924,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9367088607594937,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9240506329113924,
-      "eval_ORGANIZATION_recall": 0.9240506329113924,
+      "eval_ORGANIZATION_precision": 0.9367088607594937,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.036259282380342484,
-      "eval_overall_accuracy": 0.9884710403513588,
-      "eval_overall_f1": 0.9435382685069008,
-      "eval_overall_precision": 0.9423558897243107,
-      "eval_overall_recall": 0.9447236180904522,
-      "eval_runtime": 0.6284,
-      "eval_samples_per_second": 270.527,
-      "eval_steps_per_second": 4.774,
+      "eval_loss": 0.034257952123880386,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9522613065326633,
+      "eval_overall_precision": 0.9522613065326633,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2819,
+      "eval_samples_per_second": 603.132,
+      "eval_steps_per_second": 10.644,
       "step": 9120
     },
     {
       "epoch": 96.0,
-      "grad_norm": 0.48243898153305054,
+      "grad_norm": 0.3942119777202606,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.0306,
+      "loss": 0.0317,
       "step": 9216
     },
     {
       "epoch": 96.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.935064935064935,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9235668789808917,
+      "eval_LOCATION_precision": 0.9391304347826087,
+      "eval_LOCATION_recall": 0.9310344827586207,
+      "eval_ORGANIZATION_f1": 0.9396825396825397,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9294871794871795,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9426751592356688,
+      "eval_ORGANIZATION_recall": 0.9367088607594937,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03604614734649658,
-      "eval_overall_accuracy": 0.9881965413121054,
-      "eval_overall_f1": 0.9445843828715367,
-      "eval_overall_precision": 0.946969696969697,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6296,
-      "eval_samples_per_second": 270.033,
-      "eval_steps_per_second": 4.765,
+      "eval_loss": 0.034202940762043,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2798,
+      "eval_samples_per_second": 607.549,
+      "eval_steps_per_second": 10.721,
       "step": 9216
     },
     {
       "epoch": 97.0,
-      "grad_norm": 0.07403044402599335,
+      "grad_norm": 0.13562001287937164,
       "learning_rate": 1.5e-06,
-      "loss": 0.0304,
+      "loss": 0.0321,
       "step": 9312
     },
     {
       "epoch": 97.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.939297124600639,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9483870967741935,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.0361478365957737,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6346,
-      "eval_samples_per_second": 267.904,
-      "eval_steps_per_second": 4.728,
+      "eval_loss": 0.03407964110374451,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2822,
+      "eval_samples_per_second": 602.4,
+      "eval_steps_per_second": 10.631,
       "step": 9312
     },
     {
       "epoch": 98.0,
-      "grad_norm": 1.3072566986083984,
+      "grad_norm": 1.3050485849380493,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 0.03,
+      "loss": 0.0295,
       "step": 9408
     },
     {
       "epoch": 98.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.939297124600639,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9483870967741935,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.036295101046562195,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6248,
-      "eval_samples_per_second": 272.07,
-      "eval_steps_per_second": 4.801,
+      "eval_loss": 0.03420589491724968,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.28,
+      "eval_samples_per_second": 607.192,
+      "eval_steps_per_second": 10.715,
       "step": 9408
     },
     {
       "epoch": 99.0,
-      "grad_norm": 0.8703434467315674,
+      "grad_norm": 0.41996899247169495,
       "learning_rate": 5.000000000000001e-07,
-      "loss": 0.0293,
+      "loss": 0.031,
       "step": 9504
     },
     {
       "epoch": 99.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.939297124600639,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9483870967741935,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.036201201379299164,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6278,
-      "eval_samples_per_second": 270.799,
-      "eval_steps_per_second": 4.779,
+      "eval_loss": 0.0341360941529274,
+      "eval_overall_accuracy": 0.9898435355476256,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.282,
+      "eval_samples_per_second": 602.821,
+      "eval_steps_per_second": 10.638,
       "step": 9504
     },
     {
       "epoch": 100.0,
-      "grad_norm": 0.9420938491821289,
+      "grad_norm": 0.4611692428588867,
       "learning_rate": 0.0,
-      "loss": 0.0303,
+      "loss": 0.0299,
       "step": 9600
     },
     {
       "epoch": 100.0,
-      "eval_LOCATION_f1": 0.9264069264069263,
+      "eval_LOCATION_f1": 0.9356223175965666,
       "eval_LOCATION_number": 116,
-      "eval_LOCATION_precision": 0.9304347826086956,
-      "eval_LOCATION_recall": 0.9224137931034483,
-      "eval_ORGANIZATION_f1": 0.9206349206349206,
+      "eval_LOCATION_precision": 0.9316239316239316,
+      "eval_LOCATION_recall": 0.9396551724137931,
+      "eval_ORGANIZATION_f1": 0.939297124600639,
       "eval_ORGANIZATION_number": 158,
-      "eval_ORGANIZATION_precision": 0.9235668789808917,
-      "eval_ORGANIZATION_recall": 0.9177215189873418,
+      "eval_ORGANIZATION_precision": 0.9483870967741935,
+      "eval_ORGANIZATION_recall": 0.930379746835443,
       "eval_PERSON_f1": 0.9879518072289156,
       "eval_PERSON_number": 124,
       "eval_PERSON_precision": 0.984,
       "eval_PERSON_recall": 0.9919354838709677,
-      "eval_loss": 0.03618453070521355,
-      "eval_overall_accuracy": 0.9879220422728521,
-      "eval_overall_f1": 0.9433962264150944,
-      "eval_overall_precision": 0.9445843828715366,
-      "eval_overall_recall": 0.9422110552763819,
-      "eval_runtime": 0.6372,
-      "eval_samples_per_second": 266.78,
-      "eval_steps_per_second": 4.708,
+      "eval_loss": 0.0341842956840992,
+      "eval_overall_accuracy": 0.9895690365083722,
+      "eval_overall_f1": 0.9534591194968555,
+      "eval_overall_precision": 0.9546599496221663,
+      "eval_overall_recall": 0.9522613065326633,
+      "eval_runtime": 0.2812,
+      "eval_samples_per_second": 604.608,
+      "eval_steps_per_second": 10.67,
       "step": 9600
     },
     {
       "epoch": 100.0,
       "step": 9600,
       "total_flos": 3874712599394304.0,
-      "train_loss": 0.0784850787371397,
-      "train_runtime": 1037.5835,
-      "train_samples_per_second": 147.554,
-      "train_steps_per_second": 9.252
+      "train_loss": 0.07820479728281499,
+      "train_runtime": 511.6876,
+      "train_samples_per_second": 299.206,
+      "train_steps_per_second": 18.761
     }
   ],
   "logging_steps": 500,