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": 2.2026894092559814,
+      "grad_norm": 2.502664804458618,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 0.959,
+      "loss": 0.9512,
       "step": 106
     },
     {
@@ -37,692 +37,692 @@
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.0,
       "eval_TIME_recall": 0.0,
-      "eval_loss": 0.6148647665977478,
-      "eval_overall_accuracy": 0.8414161008729389,
+      "eval_loss": 0.6040259599685669,
+      "eval_overall_accuracy": 0.8411736178467507,
       "eval_overall_f1": 0.005714285714285714,
       "eval_overall_precision": 0.3333333333333333,
       "eval_overall_recall": 0.002881844380403458,
-      "eval_runtime": 1.0478,
-      "eval_samples_per_second": 178.467,
-      "eval_steps_per_second": 2.863,
+      "eval_runtime": 0.4042,
+      "eval_samples_per_second": 462.684,
+      "eval_steps_per_second": 7.423,
       "step": 106
     },
     {
       "epoch": 2.0,
-      "grad_norm": 1.3476005792617798,
+      "grad_norm": 1.1181972026824951,
       "learning_rate": 4.9e-05,
-      "loss": 0.5147,
+      "loss": 0.5021,
       "step": 212
     },
     {
       "epoch": 2.0,
-      "eval_LOCATION_f1": 0.4705882352941176,
+      "eval_LOCATION_f1": 0.467741935483871,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.5833333333333334,
-      "eval_LOCATION_recall": 0.39436619718309857,
-      "eval_ORGANIZATION_f1": 0.21739130434782608,
+      "eval_LOCATION_precision": 0.5471698113207547,
+      "eval_LOCATION_recall": 0.4084507042253521,
+      "eval_ORGANIZATION_f1": 0.203125,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.2054794520547945,
-      "eval_ORGANIZATION_recall": 0.23076923076923078,
-      "eval_PERSON_f1": 0.6720867208672087,
+      "eval_ORGANIZATION_precision": 0.20634920634920634,
+      "eval_ORGANIZATION_recall": 0.2,
+      "eval_PERSON_f1": 0.6997084548104956,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.5662100456621004,
-      "eval_PERSON_recall": 0.8266666666666667,
-      "eval_QUANTITY_f1": 0.21875,
+      "eval_PERSON_precision": 0.6217616580310881,
+      "eval_PERSON_recall": 0.8,
+      "eval_QUANTITY_f1": 0.1904761904761905,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.22580645161290322,
-      "eval_QUANTITY_recall": 0.21212121212121213,
-      "eval_TIME_f1": 0.5555555555555555,
+      "eval_QUANTITY_precision": 0.2,
+      "eval_QUANTITY_recall": 0.18181818181818182,
+      "eval_TIME_f1": 0.49180327868852464,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.5769230769230769,
+      "eval_TIME_precision": 0.45454545454545453,
       "eval_TIME_recall": 0.5357142857142857,
-      "eval_loss": 0.35546213388442993,
-      "eval_overall_accuracy": 0.9059165858389913,
-      "eval_overall_f1": 0.5080645161290323,
-      "eval_overall_precision": 0.4760705289672544,
-      "eval_overall_recall": 0.5446685878962536,
-      "eval_runtime": 1.0899,
-      "eval_samples_per_second": 171.576,
-      "eval_steps_per_second": 2.753,
+      "eval_loss": 0.3378148674964905,
+      "eval_overall_accuracy": 0.9076139670223085,
+      "eval_overall_f1": 0.5090403337969402,
+      "eval_overall_precision": 0.49193548387096775,
+      "eval_overall_recall": 0.5273775216138329,
+      "eval_runtime": 0.4075,
+      "eval_samples_per_second": 458.866,
+      "eval_steps_per_second": 7.361,
       "step": 212
     },
     {
       "epoch": 3.0,
-      "grad_norm": 0.8558968901634216,
+      "grad_norm": 1.0774222612380981,
       "learning_rate": 4.85e-05,
-      "loss": 0.3082,
+      "loss": 0.2912,
       "step": 318
     },
     {
       "epoch": 3.0,
-      "eval_LOCATION_f1": 0.5911949685534591,
+      "eval_LOCATION_f1": 0.6289308176100629,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.5340909090909091,
-      "eval_LOCATION_recall": 0.6619718309859155,
-      "eval_ORGANIZATION_f1": 0.5846153846153846,
+      "eval_LOCATION_precision": 0.5681818181818182,
+      "eval_LOCATION_recall": 0.704225352112676,
+      "eval_ORGANIZATION_f1": 0.5970149253731344,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5846153846153846,
-      "eval_ORGANIZATION_recall": 0.5846153846153846,
-      "eval_PERSON_f1": 0.8307692307692307,
+      "eval_ORGANIZATION_precision": 0.5797101449275363,
+      "eval_ORGANIZATION_recall": 0.6153846153846154,
+      "eval_PERSON_f1": 0.8291139240506329,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.7714285714285715,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.44999999999999996,
+      "eval_PERSON_precision": 0.7891566265060241,
+      "eval_PERSON_recall": 0.8733333333333333,
+      "eval_QUANTITY_f1": 0.5405405405405406,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.3829787234042553,
-      "eval_QUANTITY_recall": 0.5454545454545454,
+      "eval_QUANTITY_precision": 0.4878048780487805,
+      "eval_QUANTITY_recall": 0.6060606060606061,
       "eval_TIME_f1": 0.830188679245283,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.88,
       "eval_TIME_recall": 0.7857142857142857,
-      "eval_loss": 0.19337795674800873,
-      "eval_overall_accuracy": 0.9357419980601358,
-      "eval_overall_f1": 0.6961178045515395,
-      "eval_overall_precision": 0.65,
-      "eval_overall_recall": 0.7492795389048992,
-      "eval_runtime": 1.0801,
-      "eval_samples_per_second": 173.132,
-      "eval_steps_per_second": 2.778,
+      "eval_loss": 0.18643514811992645,
+      "eval_overall_accuracy": 0.9376818622696411,
+      "eval_overall_f1": 0.7146739130434783,
+      "eval_overall_precision": 0.6760925449871465,
+      "eval_overall_recall": 0.7579250720461095,
+      "eval_runtime": 0.4093,
+      "eval_samples_per_second": 456.848,
+      "eval_steps_per_second": 7.329,
       "step": 318
     },
     {
       "epoch": 4.0,
-      "grad_norm": 0.8656661510467529,
+      "grad_norm": 1.0147018432617188,
       "learning_rate": 4.8e-05,
-      "loss": 0.2009,
+      "loss": 0.1943,
       "step": 424
     },
     {
       "epoch": 4.0,
-      "eval_LOCATION_f1": 0.675,
+      "eval_LOCATION_f1": 0.7142857142857142,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6067415730337079,
-      "eval_LOCATION_recall": 0.7605633802816901,
-      "eval_ORGANIZATION_f1": 0.6114649681528661,
+      "eval_LOCATION_precision": 0.6185567010309279,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.6790123456790124,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5217391304347826,
-      "eval_ORGANIZATION_recall": 0.7384615384615385,
-      "eval_PERSON_f1": 0.8517350157728706,
+      "eval_ORGANIZATION_precision": 0.5670103092783505,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.85625,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8083832335329342,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.48101265822784817,
+      "eval_PERSON_precision": 0.8058823529411765,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.6829268292682927,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.41304347826086957,
-      "eval_QUANTITY_recall": 0.5757575757575758,
+      "eval_QUANTITY_precision": 0.5714285714285714,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.16201776266098022,
-      "eval_overall_accuracy": 0.9415615906886518,
-      "eval_overall_f1": 0.7282184655396619,
-      "eval_overall_precision": 0.6635071090047393,
-      "eval_overall_recall": 0.8069164265129684,
-      "eval_runtime": 1.0458,
-      "eval_samples_per_second": 178.803,
-      "eval_steps_per_second": 2.869,
+      "eval_loss": 0.1579989790916443,
+      "eval_overall_accuracy": 0.9478661493695442,
+      "eval_overall_f1": 0.7715736040609138,
+      "eval_overall_precision": 0.6893424036281179,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4082,
+      "eval_samples_per_second": 458.117,
+      "eval_steps_per_second": 7.349,
       "step": 424
     },
     {
       "epoch": 5.0,
-      "grad_norm": 0.9101278185844421,
+      "grad_norm": 1.6936546564102173,
       "learning_rate": 4.75e-05,
-      "loss": 0.1747,
+      "loss": 0.1653,
       "step": 530
     },
     {
       "epoch": 5.0,
-      "eval_LOCATION_f1": 0.7317073170731706,
+      "eval_LOCATION_f1": 0.7228915662650601,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6451612903225806,
+      "eval_LOCATION_precision": 0.631578947368421,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.6621621621621622,
+      "eval_ORGANIZATION_f1": 0.6712328767123288,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5903614457831325,
+      "eval_ORGANIZATION_precision": 0.6049382716049383,
       "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8698412698412699,
+      "eval_PERSON_f1": 0.8626198083067093,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8303030303030303,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7105263157894738,
+      "eval_PERSON_precision": 0.8282208588957055,
+      "eval_PERSON_recall": 0.9,
+      "eval_QUANTITY_f1": 0.8,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.627906976744186,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.7796610169491526,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7419354838709677,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14107052981853485,
-      "eval_overall_accuracy": 0.9510184287099903,
-      "eval_overall_f1": 0.7769028871391075,
-      "eval_overall_precision": 0.7132530120481928,
-      "eval_overall_recall": 0.8530259365994236,
-      "eval_runtime": 1.0459,
-      "eval_samples_per_second": 178.8,
-      "eval_steps_per_second": 2.868,
+      "eval_TIME_precision": 0.75,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.13640257716178894,
+      "eval_overall_accuracy": 0.9524733268671193,
+      "eval_overall_f1": 0.7842105263157895,
+      "eval_overall_precision": 0.7215496368038741,
+      "eval_overall_recall": 0.8587896253602305,
+      "eval_runtime": 0.4066,
+      "eval_samples_per_second": 459.951,
+      "eval_steps_per_second": 7.379,
       "step": 530
     },
     {
       "epoch": 6.0,
-      "grad_norm": 1.2041453123092651,
+      "grad_norm": 1.5403629541397095,
       "learning_rate": 4.7e-05,
-      "loss": 0.1564,
+      "loss": 0.152,
       "step": 636
     },
     {
       "epoch": 6.0,
-      "eval_LOCATION_f1": 0.7093023255813954,
+      "eval_LOCATION_f1": 0.7272727272727273,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6039603960396039,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7037037037037037,
+      "eval_LOCATION_precision": 0.6382978723404256,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.6666666666666667,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5876288659793815,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8359133126934984,
+      "eval_ORGANIZATION_precision": 0.5567010309278351,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8553846153846154,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.7803468208092486,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.6829268292682927,
+      "eval_PERSON_precision": 0.7942857142857143,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7088607594936709,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.5714285714285714,
+      "eval_QUANTITY_precision": 0.6086956521739131,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7164179104477612,
+      "eval_TIME_f1": 0.6969696969696969,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6153846153846154,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1720166802406311,
-      "eval_overall_accuracy": 0.9401066925315228,
-      "eval_overall_f1": 0.7568238213399504,
-      "eval_overall_precision": 0.664488017429194,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0496,
-      "eval_samples_per_second": 178.156,
-      "eval_steps_per_second": 2.858,
+      "eval_TIME_precision": 0.6052631578947368,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.15786363184452057,
+      "eval_overall_accuracy": 0.944713870029098,
+      "eval_overall_f1": 0.7628607277289837,
+      "eval_overall_precision": 0.6755555555555556,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.408,
+      "eval_samples_per_second": 458.337,
+      "eval_steps_per_second": 7.353,
       "step": 636
     },
     {
       "epoch": 7.0,
-      "grad_norm": 2.570847272872925,
+      "grad_norm": 2.2046146392822266,
       "learning_rate": 4.6500000000000005e-05,
-      "loss": 0.1441,
+      "loss": 0.1379,
       "step": 742
     },
     {
       "epoch": 7.0,
-      "eval_LOCATION_f1": 0.7100591715976331,
+      "eval_LOCATION_f1": 0.7439024390243902,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6122448979591837,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7080745341614907,
+      "eval_LOCATION_precision": 0.6559139784946236,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7012987012987013,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.59375,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8466257668711655,
+      "eval_ORGANIZATION_precision": 0.6067415730337079,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8553846153846154,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.7840909090909091,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.736842105263158,
+      "eval_PERSON_precision": 0.7942857142857143,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.6341463414634148,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6511627906976745,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7796610169491526,
+      "eval_QUANTITY_precision": 0.5306122448979592,
+      "eval_QUANTITY_recall": 0.7878787878787878,
+      "eval_TIME_f1": 0.7076923076923075,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7419354838709677,
+      "eval_TIME_precision": 0.6216216216216216,
       "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14682450890541077,
-      "eval_overall_accuracy": 0.9476236663433559,
-      "eval_overall_f1": 0.7737041719342604,
-      "eval_overall_precision": 0.6891891891891891,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0693,
-      "eval_samples_per_second": 174.879,
-      "eval_steps_per_second": 2.806,
+      "eval_loss": 0.15319040417671204,
+      "eval_overall_accuracy": 0.9464112512124151,
+      "eval_overall_f1": 0.7670886075949368,
+      "eval_overall_precision": 0.6839729119638827,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4079,
+      "eval_samples_per_second": 458.417,
+      "eval_steps_per_second": 7.354,
       "step": 742
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.648723304271698,
+      "grad_norm": 0.8930031061172485,
       "learning_rate": 4.600000000000001e-05,
-      "loss": 0.1299,
+      "loss": 0.125,
       "step": 848
     },
     {
       "epoch": 8.0,
-      "eval_LOCATION_f1": 0.7484662576687117,
+      "eval_LOCATION_f1": 0.7425149700598803,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6630434782608695,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7236842105263158,
+      "eval_LOCATION_precision": 0.6458333333333334,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.6962025316455697,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.632183908045977,
+      "eval_ORGANIZATION_precision": 0.5913978494623656,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8761904761904761,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8363636363636363,
+      "eval_PERSON_precision": 0.8571428571428571,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_QUANTITY_f1": 0.7605633802816901,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7105263157894737,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.12678906321525574,
-      "eval_overall_accuracy": 0.9568380213385063,
-      "eval_overall_f1": 0.8078947368421052,
-      "eval_overall_precision": 0.7433414043583535,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0505,
-      "eval_samples_per_second": 178.011,
-      "eval_steps_per_second": 2.856,
+      "eval_TIME_precision": 0.75,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.12710979580879211,
+      "eval_overall_accuracy": 0.9561105722599418,
+      "eval_overall_f1": 0.7979139504563234,
+      "eval_overall_precision": 0.7285714285714285,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4046,
+      "eval_samples_per_second": 462.173,
+      "eval_steps_per_second": 7.415,
       "step": 848
     },
     {
       "epoch": 9.0,
-      "grad_norm": 1.1211426258087158,
+      "grad_norm": 1.0148112773895264,
       "learning_rate": 4.55e-05,
-      "loss": 0.1221,
+      "loss": 0.1161,
       "step": 954
     },
     {
       "epoch": 9.0,
-      "eval_LOCATION_f1": 0.7547169811320754,
+      "eval_LOCATION_f1": 0.7320261437908497,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6818181818181818,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7248322147651007,
+      "eval_LOCATION_precision": 0.6829268292682927,
+      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_ORGANIZATION_f1": 0.6707317073170732,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6428571428571429,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8717948717948718,
+      "eval_ORGANIZATION_precision": 0.5555555555555556,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8846153846153846,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8395061728395061,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7567567567567567,
+      "eval_PERSON_precision": 0.8518518518518519,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6829268292682927,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.11980857700109482,
-      "eval_overall_accuracy": 0.9565955383123181,
-      "eval_overall_f1": 0.8021248339973439,
-      "eval_overall_precision": 0.7438423645320197,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0508,
-      "eval_samples_per_second": 177.952,
-      "eval_steps_per_second": 2.855,
+      "eval_loss": 0.12680365145206451,
+      "eval_overall_accuracy": 0.954898157129001,
+      "eval_overall_f1": 0.793148880105402,
+      "eval_overall_precision": 0.7305825242718447,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4068,
+      "eval_samples_per_second": 459.728,
+      "eval_steps_per_second": 7.375,
       "step": 954
     },
     {
       "epoch": 10.0,
-      "grad_norm": 1.1789932250976562,
+      "grad_norm": 1.5242857933044434,
       "learning_rate": 4.5e-05,
-      "loss": 0.1176,
+      "loss": 0.1124,
       "step": 1060
     },
     {
       "epoch": 10.0,
-      "eval_LOCATION_f1": 0.7169811320754716,
+      "eval_LOCATION_f1": 0.7532467532467533,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6477272727272727,
-      "eval_LOCATION_recall": 0.8028169014084507,
-      "eval_ORGANIZATION_f1": 0.7552447552447553,
+      "eval_LOCATION_precision": 0.6987951807228916,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7272727272727272,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6923076923076923,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8616352201257862,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8860759493670886,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8154761904761905,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7272727272727273,
+      "eval_PERSON_precision": 0.8433734939759037,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7567567567567567,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6363636363636364,
+      "eval_QUANTITY_precision": 0.6829268292682927,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.7868852459016394,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.7272727272727273,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.12928320467472076,
-      "eval_overall_accuracy": 0.954898157129001,
-      "eval_overall_f1": 0.7957559681697612,
-      "eval_overall_precision": 0.7371007371007371,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 1.0405,
-      "eval_samples_per_second": 179.724,
-      "eval_steps_per_second": 2.883,
+      "eval_loss": 0.12313192337751389,
+      "eval_overall_accuracy": 0.9592628516003879,
+      "eval_overall_f1": 0.8074866310160428,
+      "eval_overall_precision": 0.7531172069825436,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.4064,
+      "eval_samples_per_second": 460.091,
+      "eval_steps_per_second": 7.381,
       "step": 1060
     },
     {
       "epoch": 11.0,
-      "grad_norm": 1.4546157121658325,
+      "grad_norm": 0.9723674058914185,
       "learning_rate": 4.4500000000000004e-05,
-      "loss": 0.1111,
+      "loss": 0.1059,
       "step": 1166
     },
     {
       "epoch": 11.0,
-      "eval_LOCATION_f1": 0.7341772151898733,
+      "eval_LOCATION_f1": 0.7532467532467533,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6666666666666666,
+      "eval_LOCATION_precision": 0.6987951807228916,
       "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7323943661971831,
+      "eval_ORGANIZATION_f1": 0.7142857142857142,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6753246753246753,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8846153846153846,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8518518518518519,
+      "eval_PERSON_precision": 0.8625,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7796610169491526,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.7868852459016394,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7419354838709677,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.12095968425273895,
-      "eval_overall_accuracy": 0.9587778855480117,
-      "eval_overall_f1": 0.8048452220726784,
-      "eval_overall_precision": 0.7550505050505051,
-      "eval_overall_recall": 0.861671469740634,
-      "eval_runtime": 1.1365,
-      "eval_samples_per_second": 164.542,
-      "eval_steps_per_second": 2.64,
+      "eval_TIME_precision": 0.7272727272727273,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.1203012615442276,
+      "eval_overall_accuracy": 0.9592628516003879,
+      "eval_overall_f1": 0.8048780487804879,
+      "eval_overall_precision": 0.7595907928388747,
+      "eval_overall_recall": 0.8559077809798271,
+      "eval_runtime": 0.4057,
+      "eval_samples_per_second": 460.934,
+      "eval_steps_per_second": 7.395,
       "step": 1166
     },
     {
       "epoch": 12.0,
-      "grad_norm": 1.1820521354675293,
+      "grad_norm": 1.34227454662323,
       "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.1026,
+      "loss": 0.0975,
       "step": 1272
     },
     {
       "epoch": 12.0,
-      "eval_LOCATION_f1": 0.7239263803680982,
+      "eval_LOCATION_f1": 0.7341772151898733,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6413043478260869,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7261146496815286,
+      "eval_LOCATION_precision": 0.6666666666666666,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.6956521739130436,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6195652173913043,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.5833333333333334,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.86875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7466666666666666,
+      "eval_PERSON_precision": 0.8176470588235294,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6666666666666666,
+      "eval_QUANTITY_precision": 0.7368421052631579,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7796610169491526,
+      "eval_TIME_f1": 0.78125,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7419354838709677,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14555691182613373,
-      "eval_overall_accuracy": 0.9539282250242483,
-      "eval_overall_f1": 0.7947712418300654,
-      "eval_overall_precision": 0.7272727272727273,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0483,
-      "eval_samples_per_second": 178.389,
-      "eval_steps_per_second": 2.862,
+      "eval_TIME_precision": 0.6944444444444444,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1370622217655182,
+      "eval_overall_accuracy": 0.9527158098933075,
+      "eval_overall_f1": 0.7906976744186047,
+      "eval_overall_precision": 0.7166276346604216,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4061,
+      "eval_samples_per_second": 460.527,
+      "eval_steps_per_second": 7.388,
       "step": 1272
     },
     {
       "epoch": 13.0,
-      "grad_norm": 0.8939817547798157,
+      "grad_norm": 0.8592511415481567,
       "learning_rate": 4.35e-05,
-      "loss": 0.0958,
+      "loss": 0.0915,
       "step": 1378
     },
     {
       "epoch": 13.0,
-      "eval_LOCATION_f1": 0.7417218543046357,
+      "eval_LOCATION_f1": 0.7712418300653594,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7,
-      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_LOCATION_precision": 0.7195121951219512,
+      "eval_LOCATION_recall": 0.8309859154929577,
       "eval_ORGANIZATION_f1": 0.7199999999999999,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6352941176470588,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8817891373801918,
+      "eval_PERSON_f1": 0.8952380952380952,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8466257668711656,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8545454545454545,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7605633802816901,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_precision": 0.7105263157894737,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_TIME_f1": 0.9090909090909091,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.12335564196109772,
-      "eval_overall_accuracy": 0.957807953443259,
-      "eval_overall_f1": 0.8070175438596491,
-      "eval_overall_precision": 0.7588832487309645,
-      "eval_overall_recall": 0.861671469740634,
-      "eval_runtime": 1.0637,
-      "eval_samples_per_second": 175.806,
-      "eval_steps_per_second": 2.82,
+      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.12159363925457001,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8225806451612904,
+      "eval_overall_precision": 0.7707808564231738,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4059,
+      "eval_samples_per_second": 460.692,
+      "eval_steps_per_second": 7.391,
       "step": 1378
     },
     {
       "epoch": 14.0,
-      "grad_norm": 1.7542805671691895,
+      "grad_norm": 3.151130437850952,
       "learning_rate": 4.3e-05,
-      "loss": 0.0963,
+      "loss": 0.0913,
       "step": 1484
     },
     {
       "epoch": 14.0,
-      "eval_LOCATION_f1": 0.7662337662337662,
+      "eval_LOCATION_f1": 0.76,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7108433734939759,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7397260273972603,
+      "eval_LOCATION_precision": 0.7215189873417721,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8846153846153846,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8518518518518519,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.888888888888889,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.9230769230769231,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.11888094246387482,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8238482384823849,
-      "eval_overall_precision": 0.7774936061381074,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0551,
-      "eval_samples_per_second": 177.237,
-      "eval_steps_per_second": 2.843,
+      "eval_loss": 0.11676321178674698,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.8335625859697388,
+      "eval_overall_precision": 0.7973684210526316,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4073,
+      "eval_samples_per_second": 459.097,
+      "eval_steps_per_second": 7.365,
       "step": 1484
     },
     {
       "epoch": 15.0,
-      "grad_norm": 1.1723530292510986,
+      "grad_norm": 1.3836969137191772,
       "learning_rate": 4.25e-05,
-      "loss": 0.0877,
+      "loss": 0.0853,
       "step": 1590
     },
     {
       "epoch": 15.0,
-      "eval_LOCATION_f1": 0.7307692307692307,
+      "eval_LOCATION_f1": 0.794701986754967,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6705882352941176,
-      "eval_LOCATION_recall": 0.8028169014084507,
-      "eval_ORGANIZATION_f1": 0.728476821192053,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.738255033557047,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6395348837209303,
+      "eval_ORGANIZATION_precision": 0.6547619047619048,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_precision": 0.8571428571428571,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_QUANTITY_f1": 0.7837837837837839,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7073170731707317,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.13062922656536102,
-      "eval_overall_accuracy": 0.9565955383123181,
-      "eval_overall_f1": 0.8080536912751677,
-      "eval_overall_precision": 0.7562814070351759,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.0529,
-      "eval_samples_per_second": 177.608,
-      "eval_steps_per_second": 2.849,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.12170998007059097,
+      "eval_overall_accuracy": 0.9604752667313288,
+      "eval_overall_f1": 0.8241610738255033,
+      "eval_overall_precision": 0.7713567839195979,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.412,
+      "eval_samples_per_second": 453.856,
+      "eval_steps_per_second": 7.281,
       "step": 1590
     },
     {
       "epoch": 16.0,
-      "grad_norm": 1.6313313245773315,
+      "grad_norm": 1.1944974660873413,
       "learning_rate": 4.2e-05,
-      "loss": 0.0856,
+      "loss": 0.0809,
       "step": 1696
     },
     {
       "epoch": 16.0,
-      "eval_LOCATION_f1": 0.759493670886076,
+      "eval_LOCATION_f1": 0.782051282051282,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6896551724137931,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7432432432432433,
+      "eval_LOCATION_precision": 0.7176470588235294,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7516778523489933,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6626506024096386,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8789808917197452,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.8853503184713376,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8414634146341463,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.6923076923076923,
+      "eval_PERSON_precision": 0.8475609756097561,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7200000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6,
+      "eval_QUANTITY_precision": 0.6428571428571429,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.7796610169491526,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7419354838709677,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.1443340927362442,
-      "eval_overall_accuracy": 0.9546556741028128,
-      "eval_overall_f1": 0.8005284015852047,
-      "eval_overall_precision": 0.7390243902439024,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0604,
-      "eval_samples_per_second": 176.347,
-      "eval_steps_per_second": 2.829,
+      "eval_TIME_precision": 0.8064516129032258,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.13052447140216827,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8180610889774237,
+      "eval_overall_precision": 0.7586206896551724,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4077,
+      "eval_samples_per_second": 458.719,
+      "eval_steps_per_second": 7.359,
       "step": 1696
     },
     {
       "epoch": 17.0,
-      "grad_norm": 0.8303266167640686,
+      "grad_norm": 1.1874595880508423,
       "learning_rate": 4.15e-05,
-      "loss": 0.0811,
+      "loss": 0.0773,
       "step": 1802
     },
     {
       "epoch": 17.0,
-      "eval_LOCATION_f1": 0.7499999999999999,
+      "eval_LOCATION_f1": 0.7692307692307694,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6741573033707865,
+      "eval_LOCATION_precision": 0.7058823529411765,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7417218543046358,
+      "eval_ORGANIZATION_f1": 0.7058823529411765,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6511627906976745,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.6136363636363636,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8980891719745222,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7605633802816901,
+      "eval_PERSON_precision": 0.8597560975609756,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7105263157894737,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.819672131147541,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.13409367203712463,
-      "eval_overall_accuracy": 0.9568380213385063,
-      "eval_overall_f1": 0.8095872170439414,
-      "eval_overall_precision": 0.7524752475247525,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0536,
-      "eval_samples_per_second": 177.488,
-      "eval_steps_per_second": 2.847,
+      "eval_TIME_precision": 0.7575757575757576,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.12760519981384277,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8148148148148148,
+      "eval_overall_precision": 0.7530562347188264,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4077,
+      "eval_samples_per_second": 458.725,
+      "eval_steps_per_second": 7.359,
       "step": 1802
     },
     {
       "epoch": 18.0,
-      "grad_norm": 0.8499922752380371,
+      "grad_norm": 1.0034806728363037,
       "learning_rate": 4.1e-05,
-      "loss": 0.0779,
+      "loss": 0.0737,
       "step": 1908
     },
     {
       "epoch": 18.0,
-      "eval_LOCATION_f1": 0.7421383647798742,
+      "eval_LOCATION_f1": 0.7407407407407407,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6704545454545454,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7402597402597403,
+      "eval_LOCATION_precision": 0.6593406593406593,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.6967741935483871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6404494382022472,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8945686900958466,
+      "eval_ORGANIZATION_precision": 0.6,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8840125391849529,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8588957055214724,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.6842105263157895,
+      "eval_PERSON_precision": 0.834319526627219,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7105263157894738,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6046511627906976,
-      "eval_QUANTITY_recall": 0.7878787878787878,
-      "eval_TIME_f1": 0.7540983606557378,
+      "eval_QUANTITY_precision": 0.627906976744186,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.7796610169491526,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.696969696969697,
+      "eval_TIME_precision": 0.7419354838709677,
       "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14901402592658997,
-      "eval_overall_accuracy": 0.9529582929194956,
-      "eval_overall_f1": 0.799475753604194,
-      "eval_overall_precision": 0.7331730769230769,
+      "eval_loss": 0.15325599908828735,
+      "eval_overall_accuracy": 0.9493210475266731,
+      "eval_overall_f1": 0.7911802853437095,
+      "eval_overall_precision": 0.7193396226415094,
       "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0609,
-      "eval_samples_per_second": 176.262,
-      "eval_steps_per_second": 2.828,
+      "eval_runtime": 0.4065,
+      "eval_samples_per_second": 460.02,
+      "eval_steps_per_second": 7.38,
       "step": 1908
     },
     {
       "epoch": 19.0,
-      "grad_norm": 1.0185816287994385,
+      "grad_norm": 1.108689785003662,
       "learning_rate": 4.05e-05,
-      "loss": 0.0748,
+      "loss": 0.0703,
       "step": 2014
     },
     {
       "epoch": 19.0,
-      "eval_LOCATION_f1": 0.7516778523489933,
+      "eval_LOCATION_f1": 0.778523489932886,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.717948717948718,
-      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_LOCATION_precision": 0.7435897435897436,
+      "eval_LOCATION_recall": 0.8169014084507042,
       "eval_ORGANIZATION_f1": 0.7534246575342466,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6790123456790124,
@@ -731,3191 +731,3191 @@
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.8789808917197452,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.12444239854812622,
-      "eval_overall_accuracy": 0.9624151309408342,
-      "eval_overall_f1": 0.8296703296703296,
-      "eval_overall_precision": 0.7926509186351706,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0625,
-      "eval_samples_per_second": 176.0,
-      "eval_steps_per_second": 2.824,
+      "eval_loss": 0.1193038821220398,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8290013679890561,
+      "eval_overall_precision": 0.7890625,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.408,
+      "eval_samples_per_second": 458.324,
+      "eval_steps_per_second": 7.353,
       "step": 2014
     },
     {
       "epoch": 20.0,
-      "grad_norm": 1.3024747371673584,
+      "grad_norm": 1.1755281686782837,
       "learning_rate": 4e-05,
-      "loss": 0.073,
+      "loss": 0.0713,
       "step": 2120
     },
     {
       "epoch": 20.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.7922077922077921,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7832167832167831,
+      "eval_LOCATION_precision": 0.7349397590361446,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7236842105263158,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.717948717948718,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.632183908045977,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7105263157894738,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_QUANTITY_precision": 0.627906976744186,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.9259259259259259,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.12432579696178436,
-      "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8422496570644719,
-      "eval_overall_precision": 0.8036649214659686,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0561,
-      "eval_samples_per_second": 177.074,
-      "eval_steps_per_second": 2.841,
+      "eval_TIME_precision": 0.9615384615384616,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.12470389902591705,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.829530201342282,
+      "eval_overall_precision": 0.7763819095477387,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4074,
+      "eval_samples_per_second": 458.981,
+      "eval_steps_per_second": 7.363,
       "step": 2120
     },
     {
       "epoch": 21.0,
-      "grad_norm": 0.8448452949523926,
+      "grad_norm": 1.2563973665237427,
       "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.0694,
+      "loss": 0.0649,
       "step": 2226
     },
     {
       "epoch": 21.0,
-      "eval_LOCATION_f1": 0.7421383647798742,
+      "eval_LOCATION_f1": 0.7770700636942676,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6704545454545454,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7073170731707317,
+      "eval_LOCATION_precision": 0.7093023255813954,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5858585858585859,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9096774193548387,
+      "eval_ORGANIZATION_precision": 0.5894736842105263,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.88125,
-      "eval_PERSON_recall": 0.94,
-      "eval_QUANTITY_f1": 0.675,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7466666666666666,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.574468085106383,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8070175438596492,
+      "eval_QUANTITY_precision": 0.6666666666666666,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7931034482758621,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.1571197509765625,
-      "eval_overall_accuracy": 0.9522308438409312,
-      "eval_overall_f1": 0.8,
-      "eval_overall_precision": 0.7281323877068558,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.13802851736545563,
+      "eval_overall_accuracy": 0.95635305528613,
+      "eval_overall_f1": 0.80946123521682,
+      "eval_overall_precision": 0.7439613526570048,
       "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0507,
-      "eval_samples_per_second": 177.981,
-      "eval_steps_per_second": 2.855,
+      "eval_runtime": 0.4074,
+      "eval_samples_per_second": 459.06,
+      "eval_steps_per_second": 7.365,
       "step": 2226
     },
     {
       "epoch": 22.0,
-      "grad_norm": 0.6756930947303772,
+      "grad_norm": 0.8472046852111816,
       "learning_rate": 3.9000000000000006e-05,
-      "loss": 0.0655,
+      "loss": 0.0645,
       "step": 2332
     },
     {
       "epoch": 22.0,
-      "eval_LOCATION_f1": 0.7643312101910827,
+      "eval_LOCATION_f1": 0.75,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6976744186046512,
-      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_LOCATION_precision": 0.7037037037037037,
+      "eval_LOCATION_recall": 0.8028169014084507,
       "eval_ORGANIZATION_f1": 0.738255033557047,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6547619047619048,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7200000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_precision": 0.6428571428571429,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_TIME_f1": 0.8620689655172413,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1396949291229248,
-      "eval_overall_accuracy": 0.9585354025218235,
-      "eval_overall_f1": 0.8129205921938089,
-      "eval_overall_precision": 0.7626262626262627,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.046,
-      "eval_samples_per_second": 178.78,
-      "eval_steps_per_second": 2.868,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.14445021748542786,
+      "eval_overall_accuracy": 0.9587778855480117,
+      "eval_overall_f1": 0.8221024258760108,
+      "eval_overall_precision": 0.7721518987341772,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4092,
+      "eval_samples_per_second": 456.965,
+      "eval_steps_per_second": 7.331,
       "step": 2332
     },
     {
       "epoch": 23.0,
-      "grad_norm": 1.0572928190231323,
+      "grad_norm": 1.0579193830490112,
       "learning_rate": 3.85e-05,
-      "loss": 0.0623,
+      "loss": 0.0595,
       "step": 2438
     },
     {
       "epoch": 23.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.759493670886076,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
+      "eval_LOCATION_precision": 0.6896551724137931,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7619047619047619,
+      "eval_ORGANIZATION_f1": 0.7058823529411765,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6829268292682927,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6136363636363636,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7466666666666666,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6666666666666666,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8070175438596492,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.9090909090909091,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7931034482758621,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.13910295069217682,
-      "eval_overall_accuracy": 0.9590203685741998,
-      "eval_overall_f1": 0.8221024258760108,
-      "eval_overall_precision": 0.7721518987341772,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0543,
-      "eval_samples_per_second": 177.367,
-      "eval_steps_per_second": 2.845,
+      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.13738445937633514,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8284182305630027,
+      "eval_overall_precision": 0.7744360902255639,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4089,
+      "eval_samples_per_second": 457.323,
+      "eval_steps_per_second": 7.337,
       "step": 2438
     },
     {
       "epoch": 24.0,
-      "grad_norm": 0.8183007836341858,
+      "grad_norm": 1.30304753780365,
       "learning_rate": 3.8e-05,
-      "loss": 0.0588,
+      "loss": 0.0576,
       "step": 2544
     },
     {
       "epoch": 24.0,
-      "eval_LOCATION_f1": 0.7374999999999999,
+      "eval_LOCATION_f1": 0.7692307692307694,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6629213483146067,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7320261437908497,
+      "eval_LOCATION_precision": 0.7058823529411765,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7172413793103448,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6363636363636364,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.65,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.90032154340836,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.9090909090909091,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.14739328622817993,
-      "eval_overall_accuracy": 0.9575654704170709,
-      "eval_overall_f1": 0.8106666666666668,
-      "eval_overall_precision": 0.7543424317617866,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0524,
-      "eval_samples_per_second": 177.697,
-      "eval_steps_per_second": 2.851,
+      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.14021030068397522,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8276797829036635,
+      "eval_overall_precision": 0.782051282051282,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.408,
+      "eval_samples_per_second": 458.303,
+      "eval_steps_per_second": 7.352,
       "step": 2544
     },
     {
       "epoch": 25.0,
-      "grad_norm": 1.561627984046936,
+      "grad_norm": 0.7619758248329163,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0602,
+      "loss": 0.0562,
       "step": 2650
     },
     {
       "epoch": 25.0,
-      "eval_LOCATION_f1": 0.751592356687898,
+      "eval_LOCATION_f1": 0.7692307692307694,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.686046511627907,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7402597402597403,
+      "eval_LOCATION_precision": 0.7058823529411765,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.708860759493671,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6404494382022472,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_ORGANIZATION_precision": 0.6021505376344086,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
       "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.875,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7200000000000001,
+      "eval_QUANTITY_f1": 0.7105263157894738,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6428571428571429,
+      "eval_QUANTITY_precision": 0.627906976744186,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8363636363636364,
+      "eval_TIME_f1": 0.819672131147541,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8518518518518519,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14714130759239197,
-      "eval_overall_accuracy": 0.95635305528613,
-      "eval_overall_f1": 0.8149134487350199,
-      "eval_overall_precision": 0.7574257425742574,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0522,
-      "eval_samples_per_second": 177.728,
-      "eval_steps_per_second": 2.851,
+      "eval_TIME_precision": 0.7575757575757576,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.15844029188156128,
+      "eval_overall_accuracy": 0.9553831231813773,
+      "eval_overall_f1": 0.80946123521682,
+      "eval_overall_precision": 0.7439613526570048,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4076,
+      "eval_samples_per_second": 458.809,
+      "eval_steps_per_second": 7.361,
       "step": 2650
     },
     {
       "epoch": 26.0,
-      "grad_norm": 0.37913623452186584,
+      "grad_norm": 0.3145362138748169,
       "learning_rate": 3.7e-05,
-      "loss": 0.0543,
+      "loss": 0.0533,
       "step": 2756
     },
     {
       "epoch": 26.0,
-      "eval_LOCATION_f1": 0.7547169811320754,
+      "eval_LOCATION_f1": 0.7466666666666666,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6818181818181818,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.757142857142857,
+      "eval_LOCATION_precision": 0.7088607594936709,
+      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_ORGANIZATION_f1": 0.7248322147651007,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7066666666666667,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8846153846153846,
+      "eval_ORGANIZATION_precision": 0.6428571428571429,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.888888888888889,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8518518518518519,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7179487179487178,
+      "eval_PERSON_precision": 0.8484848484848485,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.6842105263157895,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6222222222222222,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.793103448275862,
+      "eval_QUANTITY_precision": 0.6046511627906976,
+      "eval_QUANTITY_recall": 0.7878787878787878,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7666666666666667,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.1607346087694168,
-      "eval_overall_accuracy": 0.9561105722599418,
-      "eval_overall_f1": 0.8085676037483265,
-      "eval_overall_precision": 0.755,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0494,
-      "eval_samples_per_second": 178.198,
-      "eval_steps_per_second": 2.859,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1501431167125702,
+      "eval_overall_accuracy": 0.9539282250242483,
+      "eval_overall_f1": 0.8058902275769746,
+      "eval_overall_precision": 0.7525,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4072,
+      "eval_samples_per_second": 459.248,
+      "eval_steps_per_second": 7.368,
       "step": 2756
     },
     {
       "epoch": 27.0,
-      "grad_norm": 1.6127158403396606,
+      "grad_norm": 1.4819648265838623,
       "learning_rate": 3.65e-05,
-      "loss": 0.0555,
+      "loss": 0.0529,
       "step": 2862
     },
     {
       "epoch": 27.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.7770700636942676,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7368421052631579,
+      "eval_LOCATION_precision": 0.7093023255813954,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7320261437908497,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6436781609195402,
+      "eval_ORGANIZATION_precision": 0.6363636363636364,
       "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_precision": 0.86875,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7200000000000001,
+      "eval_QUANTITY_f1": 0.7466666666666666,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6428571428571429,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8214285714285714,
+      "eval_QUANTITY_precision": 0.6666666666666666,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8214285714285714,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.14687323570251465,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8183041722745625,
-      "eval_overall_precision": 0.7676767676767676,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0568,
-      "eval_samples_per_second": 176.954,
-      "eval_steps_per_second": 2.839,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.15187156200408936,
+      "eval_overall_accuracy": 0.9575654704170709,
+      "eval_overall_f1": 0.8229027962716378,
+      "eval_overall_precision": 0.7648514851485149,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4068,
+      "eval_samples_per_second": 459.716,
+      "eval_steps_per_second": 7.375,
       "step": 2862
     },
     {
       "epoch": 28.0,
-      "grad_norm": 1.0716761350631714,
+      "grad_norm": 0.8129953145980835,
       "learning_rate": 3.6e-05,
-      "loss": 0.0507,
+      "loss": 0.0531,
       "step": 2968
     },
     {
       "epoch": 28.0,
-      "eval_LOCATION_f1": 0.7682119205298014,
+      "eval_LOCATION_f1": 0.802721088435374,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.725,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7295597484276729,
+      "eval_LOCATION_precision": 0.7763157894736842,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7702702702702702,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6170212765957447,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9032258064516129,
+      "eval_ORGANIZATION_precision": 0.6867469879518072,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.875,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7027027027027027,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7605633802816901,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6341463414634146,
-      "eval_QUANTITY_recall": 0.7878787878787878,
-      "eval_TIME_f1": 0.8727272727272727,
-      "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.15819337964057922,
-      "eval_overall_accuracy": 0.9546556741028128,
-      "eval_overall_f1": 0.8170894526034712,
-      "eval_overall_precision": 0.7611940298507462,
+      "eval_QUANTITY_precision": 0.7105263157894737,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.9090909090909091,
+      "eval_TIME_number": 28,
+      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.13637924194335938,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8372093023255814,
+      "eval_overall_precision": 0.796875,
       "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0494,
-      "eval_samples_per_second": 178.203,
-      "eval_steps_per_second": 2.859,
+      "eval_runtime": 0.406,
+      "eval_samples_per_second": 460.578,
+      "eval_steps_per_second": 7.389,
       "step": 2968
     },
     {
       "epoch": 29.0,
-      "grad_norm": 0.6914520263671875,
+      "grad_norm": 0.46542084217071533,
       "learning_rate": 3.55e-05,
-      "loss": 0.0512,
+      "loss": 0.0463,
       "step": 3074
     },
     {
       "epoch": 29.0,
-      "eval_LOCATION_f1": 0.7682119205298014,
+      "eval_LOCATION_f1": 0.7702702702702703,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.725,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.686046511627907,
+      "eval_LOCATION_precision": 0.7402597402597403,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5514018691588785,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.8932038834951457,
+      "eval_ORGANIZATION_precision": 0.6551724137931034,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.9038461538461539,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8679245283018868,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8703703703703703,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8405797101449276,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.8055555555555556,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.16121703386306763,
-      "eval_overall_accuracy": 0.954898157129001,
-      "eval_overall_f1": 0.8031496062992126,
-      "eval_overall_precision": 0.7373493975903614,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0541,
-      "eval_samples_per_second": 177.397,
-      "eval_steps_per_second": 2.846,
+      "eval_loss": 0.13960041105747223,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.835820895522388,
+      "eval_overall_precision": 0.7897435897435897,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4068,
+      "eval_samples_per_second": 459.668,
+      "eval_steps_per_second": 7.374,
       "step": 3074
     },
     {
       "epoch": 30.0,
-      "grad_norm": 0.9099843502044678,
+      "grad_norm": 0.8071898818016052,
       "learning_rate": 3.5e-05,
-      "loss": 0.047,
+      "loss": 0.0466,
       "step": 3180
     },
     {
       "epoch": 30.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.7919463087248323,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_LOCATION_precision": 0.7564102564102564,
+      "eval_LOCATION_recall": 0.8309859154929577,
       "eval_ORGANIZATION_f1": 0.7261146496815286,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6195652173913043,
       "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_PERSON_f1": 0.90032154340836,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7567567567567567,
+      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6829268292682927,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8363636363636364,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8518518518518519,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.1513780951499939,
-      "eval_overall_accuracy": 0.9599903006789525,
-      "eval_overall_f1": 0.8247978436657682,
-      "eval_overall_precision": 0.7746835443037975,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.1364,
-      "eval_samples_per_second": 164.552,
-      "eval_steps_per_second": 2.64,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1534765511751175,
+      "eval_overall_accuracy": 0.9585354025218235,
+      "eval_overall_f1": 0.824631860776439,
+      "eval_overall_precision": 0.77,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4052,
+      "eval_samples_per_second": 461.503,
+      "eval_steps_per_second": 7.404,
       "step": 3180
     },
     {
       "epoch": 31.0,
-      "grad_norm": 0.5555219054222107,
+      "grad_norm": 0.7920342087745667,
       "learning_rate": 3.45e-05,
-      "loss": 0.0466,
+      "loss": 0.0449,
       "step": 3286
     },
     {
       "epoch": 31.0,
-      "eval_LOCATION_f1": 0.7483870967741935,
+      "eval_LOCATION_f1": 0.7770700636942676,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6904761904761905,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7516778523489933,
+      "eval_LOCATION_precision": 0.7093023255813954,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9025974025974027,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9102564102564102,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.879746835443038,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8765432098765432,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7567567567567567,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.6829268292682927,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.9090909090909091,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9259259259259259,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.15754006803035736,
-      "eval_overall_accuracy": 0.9582929194956353,
-      "eval_overall_f1": 0.827027027027027,
-      "eval_overall_precision": 0.7786259541984732,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0609,
-      "eval_samples_per_second": 176.262,
-      "eval_steps_per_second": 2.828,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.16083230078220367,
+      "eval_overall_accuracy": 0.9592628516003879,
+      "eval_overall_f1": 0.8306451612903226,
+      "eval_overall_precision": 0.7783375314861462,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4039,
+      "eval_samples_per_second": 462.994,
+      "eval_steps_per_second": 7.428,
       "step": 3286
     },
     {
       "epoch": 32.0,
-      "grad_norm": 0.21328239142894745,
+      "grad_norm": 0.9836466908454895,
       "learning_rate": 3.4000000000000007e-05,
-      "loss": 0.0449,
+      "loss": 0.043,
       "step": 3392
     },
     {
       "epoch": 32.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.751592356687898,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7295597484276729,
+      "eval_LOCATION_precision": 0.686046511627907,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7058823529411765,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6170212765957447,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.6136363636363636,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9102564102564102,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8765432098765432,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.6933333333333334,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_QUANTITY_precision": 0.6190476190476191,
+      "eval_QUANTITY_recall": 0.7878787878787878,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.17104628682136536,
-      "eval_overall_accuracy": 0.9568380213385063,
-      "eval_overall_f1": 0.8170894526034712,
-      "eval_overall_precision": 0.7611940298507462,
+      "eval_TIME_precision": 0.8064516129032258,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.16350191831588745,
+      "eval_overall_accuracy": 0.9556256062075654,
+      "eval_overall_f1": 0.8095238095238096,
+      "eval_overall_precision": 0.7481662591687042,
       "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.2405,
-      "eval_samples_per_second": 150.742,
-      "eval_steps_per_second": 2.418,
+      "eval_runtime": 0.4041,
+      "eval_samples_per_second": 462.793,
+      "eval_steps_per_second": 7.424,
       "step": 3392
     },
     {
       "epoch": 33.0,
-      "grad_norm": 0.6623594760894775,
+      "grad_norm": 0.8012641668319702,
       "learning_rate": 3.35e-05,
-      "loss": 0.0414,
+      "loss": 0.0394,
       "step": 3498
     },
     {
       "epoch": 33.0,
-      "eval_LOCATION_f1": 0.7532467532467533,
+      "eval_LOCATION_f1": 0.7755102040816326,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6987951807228916,
-      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8028169014084507,
       "eval_ORGANIZATION_f1": 0.726027397260274,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.654320987654321,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8831168831168831,
+      "eval_PERSON_f1": 0.9220779220779222,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8607594936708861,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.8987341772151899,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
+      "eval_TIME_precision": 0.8888888888888888,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1599014401435852,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8163265306122449,
-      "eval_overall_precision": 0.7731958762886598,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 1.051,
-      "eval_samples_per_second": 177.926,
-      "eval_steps_per_second": 2.854,
+      "eval_loss": 0.15030378103256226,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8402203856749312,
+      "eval_overall_precision": 0.8047493403693932,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4047,
+      "eval_samples_per_second": 462.071,
+      "eval_steps_per_second": 7.413,
       "step": 3498
     },
     {
       "epoch": 34.0,
-      "grad_norm": 0.7860094308853149,
+      "grad_norm": 0.9052590727806091,
       "learning_rate": 3.3e-05,
-      "loss": 0.0425,
+      "loss": 0.0412,
       "step": 3604
     },
     {
       "epoch": 34.0,
-      "eval_LOCATION_f1": 0.7662337662337662,
+      "eval_LOCATION_f1": 0.7972972972972973,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7108433734939759,
+      "eval_LOCATION_precision": 0.7662337662337663,
       "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7516778523489933,
+      "eval_ORGANIZATION_f1": 0.7755102040816327,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8692810457516339,
+      "eval_ORGANIZATION_precision": 0.6951219512195121,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.9220779220779222,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8525641025641025,
-      "eval_PERSON_recall": 0.8866666666666667,
+      "eval_PERSON_precision": 0.8987341772151899,
+      "eval_PERSON_recall": 0.9466666666666667,
       "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.8888888888888888,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1540924459695816,
-      "eval_overall_accuracy": 0.9590203685741998,
-      "eval_overall_f1": 0.8108108108108107,
-      "eval_overall_precision": 0.7633587786259542,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 1.0533,
-      "eval_samples_per_second": 177.543,
-      "eval_steps_per_second": 2.848,
+      "eval_loss": 0.1466250866651535,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8481532147742818,
+      "eval_overall_precision": 0.8072916666666666,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.4032,
+      "eval_samples_per_second": 463.733,
+      "eval_steps_per_second": 7.44,
       "step": 3604
     },
     {
       "epoch": 35.0,
-      "grad_norm": 0.8806631565093994,
+      "grad_norm": 0.42126646637916565,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.0389,
+      "loss": 0.0403,
       "step": 3710
     },
     {
       "epoch": 35.0,
-      "eval_LOCATION_f1": 0.729559748427673,
+      "eval_LOCATION_f1": 0.7741935483870968,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6590909090909091,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7388535031847134,
+      "eval_LOCATION_precision": 0.7142857142857143,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7448275862068966,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6304347826086957,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8932038834951457,
+      "eval_ORGANIZATION_precision": 0.675,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9102564102564102,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8679245283018868,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.736842105263158,
+      "eval_PERSON_precision": 0.8765432098765432,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7297297297297297,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6511627906976745,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7419354838709677,
+      "eval_QUANTITY_precision": 0.6585365853658537,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6764705882352942,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.19029554724693298,
-      "eval_overall_accuracy": 0.9522308438409312,
-      "eval_overall_f1": 0.799475753604194,
-      "eval_overall_precision": 0.7331730769230769,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0532,
-      "eval_samples_per_second": 177.547,
-      "eval_steps_per_second": 2.848,
+      "eval_TIME_precision": 0.8064516129032258,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.15250608325004578,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8268456375838926,
+      "eval_overall_precision": 0.7738693467336684,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4038,
+      "eval_samples_per_second": 463.06,
+      "eval_steps_per_second": 7.429,
       "step": 3710
     },
     {
       "epoch": 36.0,
-      "grad_norm": 1.029179334640503,
+      "grad_norm": 1.7094526290893555,
       "learning_rate": 3.2000000000000005e-05,
-      "loss": 0.0389,
+      "loss": 0.0385,
       "step": 3816
     },
     {
       "epoch": 36.0,
-      "eval_LOCATION_f1": 0.7662337662337662,
+      "eval_LOCATION_f1": 0.7919463087248323,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7108433734939759,
+      "eval_LOCATION_precision": 0.7564102564102564,
       "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7169811320754718,
+      "eval_ORGANIZATION_f1": 0.7354838709677419,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6063829787234043,
+      "eval_ORGANIZATION_precision": 0.6333333333333333,
       "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.90032154340836,
+      "eval_PERSON_f1": 0.8958990536277603,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8695652173913043,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_PERSON_precision": 0.8502994011976048,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7105263157894738,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
+      "eval_QUANTITY_precision": 0.627906976744186,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.17856962978839874,
-      "eval_overall_accuracy": 0.9568380213385063,
-      "eval_overall_f1": 0.8164893617021276,
-      "eval_overall_precision": 0.7580246913580246,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0617,
-      "eval_samples_per_second": 176.128,
-      "eval_steps_per_second": 2.826,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1817101240158081,
+      "eval_overall_accuracy": 0.9551406401551892,
+      "eval_overall_f1": 0.8233731739707836,
+      "eval_overall_precision": 0.7635467980295566,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.4033,
+      "eval_samples_per_second": 463.63,
+      "eval_steps_per_second": 7.438,
       "step": 3816
     },
     {
       "epoch": 37.0,
-      "grad_norm": 1.0397053956985474,
+      "grad_norm": 1.976254940032959,
       "learning_rate": 3.15e-05,
-      "loss": 0.0369,
+      "loss": 0.037,
       "step": 3922
     },
     {
       "epoch": 37.0,
-      "eval_LOCATION_f1": 0.725,
+      "eval_LOCATION_f1": 0.7435897435897436,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.651685393258427,
+      "eval_LOCATION_precision": 0.6823529411764706,
       "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7341772151898733,
+      "eval_ORGANIZATION_f1": 0.7152317880794703,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6236559139784946,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8974358974358974,
+      "eval_ORGANIZATION_precision": 0.627906976744186,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.89937106918239,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8641975308641975,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7088607594936709,
+      "eval_PERSON_precision": 0.8511904761904762,
+      "eval_PERSON_recall": 0.9533333333333334,
+      "eval_QUANTITY_f1": 0.6582278481012658,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6086956521739131,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_QUANTITY_precision": 0.5652173913043478,
+      "eval_QUANTITY_recall": 0.7878787878787878,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.20172427594661713,
-      "eval_overall_accuracy": 0.9529582929194956,
-      "eval_overall_f1": 0.8067885117493473,
-      "eval_overall_precision": 0.7374701670644391,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 1.0533,
-      "eval_samples_per_second": 177.542,
-      "eval_steps_per_second": 2.848,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.2011863738298416,
+      "eval_overall_accuracy": 0.9517458777885548,
+      "eval_overall_f1": 0.799475753604194,
+      "eval_overall_precision": 0.7331730769230769,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4048,
+      "eval_samples_per_second": 461.918,
+      "eval_steps_per_second": 7.41,
       "step": 3922
     },
     {
       "epoch": 38.0,
-      "grad_norm": 1.4343793392181396,
+      "grad_norm": 1.1307073831558228,
       "learning_rate": 3.1e-05,
-      "loss": 0.0362,
+      "loss": 0.037,
       "step": 4028
     },
     {
       "epoch": 38.0,
-      "eval_LOCATION_f1": 0.7612903225806451,
+      "eval_LOCATION_f1": 0.7837837837837838,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7023809523809523,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7464788732394366,
+      "eval_LOCATION_precision": 0.7532467532467533,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.9259259259259259,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1783708781003952,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8141112618724559,
-      "eval_overall_precision": 0.7692307692307693,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 1.0559,
-      "eval_samples_per_second": 177.107,
-      "eval_steps_per_second": 2.841,
+      "eval_TIME_precision": 0.9615384615384616,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1582067906856537,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.8381742738589211,
+      "eval_overall_precision": 0.8058510638297872,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4045,
+      "eval_samples_per_second": 462.274,
+      "eval_steps_per_second": 7.416,
       "step": 4028
     },
     {
       "epoch": 39.0,
-      "grad_norm": 0.2365792840719223,
+      "grad_norm": 0.8394156694412231,
       "learning_rate": 3.05e-05,
-      "loss": 0.034,
+      "loss": 0.0332,
       "step": 4134
     },
     {
       "epoch": 39.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.794701986754967,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
+      "eval_LOCATION_precision": 0.75,
       "eval_LOCATION_recall": 0.8450704225352113,
       "eval_ORGANIZATION_f1": 0.76,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6705882352941176,
       "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_PERSON_f1": 0.9108280254777071,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_PERSON_precision": 0.8719512195121951,
+      "eval_PERSON_recall": 0.9533333333333334,
       "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.675,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.17429344356060028,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8265582655826558,
-      "eval_overall_precision": 0.7800511508951407,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0434,
-      "eval_samples_per_second": 179.224,
-      "eval_steps_per_second": 2.875,
+      "eval_loss": 0.16989700496196747,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8360215053763441,
+      "eval_overall_precision": 0.783375314861461,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.4173,
+      "eval_samples_per_second": 448.072,
+      "eval_steps_per_second": 7.188,
       "step": 4134
     },
     {
       "epoch": 40.0,
-      "grad_norm": 1.8789796829223633,
+      "grad_norm": 0.8886793255805969,
       "learning_rate": 3e-05,
-      "loss": 0.0342,
+      "loss": 0.0354,
       "step": 4240
     },
     {
       "epoch": 40.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.794701986754967,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
+      "eval_LOCATION_precision": 0.75,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7702702702702702,
+      "eval_ORGANIZATION_f1": 0.7412587412587412,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6867469879518072,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8938906752411575,
+      "eval_ORGANIZATION_precision": 0.6794871794871795,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9073482428115016,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8633540372670807,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_PERSON_precision": 0.8711656441717791,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_precision": 0.7,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.7741935483870968,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7058823529411765,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1677405685186386,
-      "eval_overall_accuracy": 0.9624151309408342,
-      "eval_overall_f1": 0.8335588633288228,
-      "eval_overall_precision": 0.7857142857142857,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0476,
-      "eval_samples_per_second": 178.511,
-      "eval_steps_per_second": 2.864,
+      "eval_loss": 0.1586233377456665,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.8274932614555257,
+      "eval_overall_precision": 0.7772151898734178,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4013,
+      "eval_samples_per_second": 465.977,
+      "eval_steps_per_second": 7.476,
       "step": 4240
     },
     {
       "epoch": 41.0,
-      "grad_norm": 0.3054400086402893,
+      "grad_norm": 0.6773855686187744,
       "learning_rate": 2.95e-05,
-      "loss": 0.0325,
+      "loss": 0.0331,
       "step": 4346
     },
     {
       "epoch": 41.0,
-      "eval_LOCATION_f1": 0.759493670886076,
+      "eval_LOCATION_f1": 0.7741935483870968,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6896551724137931,
+      "eval_LOCATION_precision": 0.7142857142857143,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7581699346405228,
+      "eval_ORGANIZATION_f1": 0.7619047619047619,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6590909090909091,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.887459807073955,
+      "eval_ORGANIZATION_precision": 0.6829268292682927,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8571428571428571,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7466666666666666,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6666666666666666,
+      "eval_QUANTITY_precision": 0.7,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19720277190208435,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.816976127320955,
-      "eval_overall_precision": 0.7567567567567568,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0527,
-      "eval_samples_per_second": 177.643,
-      "eval_steps_per_second": 2.85,
+      "eval_loss": 0.16328278183937073,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.8286099865047233,
+      "eval_overall_precision": 0.7791878172588832,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4108,
+      "eval_samples_per_second": 455.24,
+      "eval_steps_per_second": 7.303,
       "step": 4346
     },
     {
       "epoch": 42.0,
-      "grad_norm": 0.7684195637702942,
+      "grad_norm": 0.5810602903366089,
       "learning_rate": 2.9e-05,
-      "loss": 0.0304,
+      "loss": 0.0312,
       "step": 4452
     },
     {
       "epoch": 42.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.778523489932886,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_LOCATION_precision": 0.7435897435897436,
+      "eval_LOCATION_recall": 0.8169014084507042,
       "eval_ORGANIZATION_f1": 0.7448275862068966,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.675,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_PERSON_f1": 0.9009584664536742,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_PERSON_precision": 0.8650306748466258,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7123287671232875,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.65,
+      "eval_QUANTITY_recall": 0.7878787878787878,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.18190883100032806,
-      "eval_overall_accuracy": 0.9585354025218235,
-      "eval_overall_f1": 0.8184281842818427,
-      "eval_overall_precision": 0.7723785166240409,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0594,
-      "eval_samples_per_second": 176.508,
-      "eval_steps_per_second": 2.832,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1705748438835144,
+      "eval_overall_accuracy": 0.9595053346265762,
+      "eval_overall_f1": 0.824966078697422,
+      "eval_overall_precision": 0.7794871794871795,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4042,
+      "eval_samples_per_second": 462.62,
+      "eval_steps_per_second": 7.422,
       "step": 4452
     },
     {
       "epoch": 43.0,
-      "grad_norm": 0.31616175174713135,
+      "grad_norm": 1.03245210647583,
       "learning_rate": 2.8499999999999998e-05,
-      "loss": 0.0295,
+      "loss": 0.031,
       "step": 4558
     },
     {
       "epoch": 43.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.782051282051282,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7516778523489933,
+      "eval_LOCATION_precision": 0.7176470588235294,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7482014388489208,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.7027027027027027,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9073482428115016,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7105263157894738,
+      "eval_PERSON_precision": 0.8711656441717791,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.821917808219178,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.627906976744186,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_QUANTITY_precision": 0.75,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19346393644809723,
-      "eval_overall_accuracy": 0.9575654704170709,
-      "eval_overall_f1": 0.8183041722745625,
-      "eval_overall_precision": 0.7676767676767676,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0521,
-      "eval_samples_per_second": 177.732,
-      "eval_steps_per_second": 2.851,
+      "eval_loss": 0.16453760862350464,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8362652232746954,
+      "eval_overall_precision": 0.7882653061224489,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4028,
+      "eval_samples_per_second": 464.207,
+      "eval_steps_per_second": 7.447,
       "step": 4558
     },
     {
       "epoch": 44.0,
-      "grad_norm": 2.207120418548584,
+      "grad_norm": 1.3175857067108154,
       "learning_rate": 2.8000000000000003e-05,
-      "loss": 0.0301,
+      "loss": 0.0305,
       "step": 4664
     },
     {
       "epoch": 44.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.7843137254901961,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7307692307692307,
+      "eval_LOCATION_precision": 0.7317073170731707,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7549668874172185,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6263736263736264,
+      "eval_ORGANIZATION_precision": 0.6627906976744186,
       "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9061488673139159,
+      "eval_PERSON_f1": 0.9044585987261147,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8805031446540881,
-      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_PERSON_precision": 0.8658536585365854,
+      "eval_PERSON_recall": 0.9466666666666667,
       "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.675,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19700577855110168,
-      "eval_overall_accuracy": 0.957807953443259,
-      "eval_overall_f1": 0.8306451612903226,
-      "eval_overall_precision": 0.7783375314861462,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 1.0474,
-      "eval_samples_per_second": 178.541,
-      "eval_steps_per_second": 2.864,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1853325217962265,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8326639892904953,
+      "eval_overall_precision": 0.7775,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.4048,
+      "eval_samples_per_second": 462.013,
+      "eval_steps_per_second": 7.412,
       "step": 4664
     },
     {
       "epoch": 45.0,
-      "grad_norm": 0.8179439902305603,
+      "grad_norm": 2.4794533252716064,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.0291,
+      "loss": 0.0284,
       "step": 4770
     },
     {
       "epoch": 45.0,
-      "eval_LOCATION_f1": 0.782051282051282,
+      "eval_LOCATION_f1": 0.7866666666666666,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7176470588235294,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7862068965517242,
+      "eval_LOCATION_precision": 0.7468354430379747,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7125,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8903225806451613,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8625,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19478309154510498,
-      "eval_overall_accuracy": 0.9595053346265762,
-      "eval_overall_f1": 0.8306451612903226,
-      "eval_overall_precision": 0.7783375314861462,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 1.1283,
-      "eval_samples_per_second": 165.73,
-      "eval_steps_per_second": 2.659,
+      "eval_loss": 0.16584299504756927,
+      "eval_overall_accuracy": 0.9633850630455868,
+      "eval_overall_f1": 0.8308115543328748,
+      "eval_overall_precision": 0.7947368421052632,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.4049,
+      "eval_samples_per_second": 461.854,
+      "eval_steps_per_second": 7.409,
       "step": 4770
     },
     {
       "epoch": 46.0,
-      "grad_norm": 0.5856307148933411,
+      "grad_norm": 1.2760629653930664,
       "learning_rate": 2.7000000000000002e-05,
-      "loss": 0.0274,
+      "loss": 0.028,
       "step": 4876
     },
     {
       "epoch": 46.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.7581699346405228,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7785234899328859,
+      "eval_LOCATION_precision": 0.7073170731707317,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7210884353741497,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6904761904761905,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8910891089108911,
+      "eval_ORGANIZATION_precision": 0.6463414634146342,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8823529411764706,
-      "eval_PERSON_recall": 0.9,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
       "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.675,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.18428590893745422,
-      "eval_overall_accuracy": 0.960717749757517,
-      "eval_overall_f1": 0.8299319727891157,
-      "eval_overall_precision": 0.7860824742268041,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0545,
-      "eval_samples_per_second": 177.342,
-      "eval_steps_per_second": 2.845,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.17332231998443604,
+      "eval_overall_accuracy": 0.9599903006789525,
+      "eval_overall_f1": 0.8238482384823849,
+      "eval_overall_precision": 0.7774936061381074,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4033,
+      "eval_samples_per_second": 463.665,
+      "eval_steps_per_second": 7.438,
       "step": 4876
     },
     {
       "epoch": 47.0,
-      "grad_norm": 0.8249619007110596,
+      "grad_norm": 0.9426093101501465,
       "learning_rate": 2.6500000000000004e-05,
-      "loss": 0.0249,
+      "loss": 0.0257,
       "step": 4982
     },
     {
       "epoch": 47.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.802721088435374,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7619047619047619,
+      "eval_LOCATION_precision": 0.7763157894736842,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7368421052631579,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6829268292682927,
+      "eval_ORGANIZATION_precision": 0.6436781609195402,
       "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8846153846153846,
+      "eval_PERSON_f1": 0.9073482428115016,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8518518518518519,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7105263157894738,
+      "eval_PERSON_precision": 0.8711656441717791,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.627906976744186,
+      "eval_QUANTITY_precision": 0.6923076923076923,
       "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2020081728696823,
-      "eval_overall_accuracy": 0.9587778855480117,
-      "eval_overall_f1": 0.8187919463087249,
-      "eval_overall_precision": 0.7663316582914573,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0517,
-      "eval_samples_per_second": 177.806,
-      "eval_steps_per_second": 2.853,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.18327540159225464,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.834008097165992,
+      "eval_overall_precision": 0.7842639593908629,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4078,
+      "eval_samples_per_second": 458.511,
+      "eval_steps_per_second": 7.356,
       "step": 4982
     },
     {
       "epoch": 48.0,
-      "grad_norm": 1.4953243732452393,
+      "grad_norm": 0.6249455213546753,
       "learning_rate": 2.6000000000000002e-05,
-      "loss": 0.0259,
+      "loss": 0.0281,
       "step": 5088
     },
     {
       "epoch": 48.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.7837837837837838,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7857142857142856,
+      "eval_LOCATION_precision": 0.7532467532467533,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7310344827586207,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7333333333333333,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.6625,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8910256410256411,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8580246913580247,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.8518518518518519,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.18837600946426392,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8333333333333333,
-      "eval_overall_precision": 0.7922077922077922,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0473,
-      "eval_samples_per_second": 178.558,
-      "eval_steps_per_second": 2.865,
+      "eval_TIME_precision": 0.8846153846153846,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.17024332284927368,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.823529411764706,
+      "eval_overall_precision": 0.7838541666666666,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4064,
+      "eval_samples_per_second": 460.192,
+      "eval_steps_per_second": 7.383,
       "step": 5088
     },
     {
       "epoch": 49.0,
-      "grad_norm": 0.07555894553661346,
+      "grad_norm": 0.06978233903646469,
       "learning_rate": 2.5500000000000003e-05,
-      "loss": 0.0252,
+      "loss": 0.0245,
       "step": 5194
     },
     {
       "epoch": 49.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7741935483870968,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7746478873239436,
+      "eval_LOCATION_precision": 0.7142857142857143,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7346938775510204,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6585365853658537,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9131832797427653,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.821917808219178,
+      "eval_PERSON_precision": 0.8819875776397516,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.888888888888889,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8620689655172413,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9230769230769231,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.18851637840270996,
-      "eval_overall_accuracy": 0.9626576139670223,
-      "eval_overall_f1": 0.8410958904109589,
-      "eval_overall_precision": 0.8015665796344648,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0481,
-      "eval_samples_per_second": 178.415,
-      "eval_steps_per_second": 2.862,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.18626831471920013,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8279569892473119,
+      "eval_overall_precision": 0.7758186397984886,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4028,
+      "eval_samples_per_second": 464.25,
+      "eval_steps_per_second": 7.448,
       "step": 5194
     },
     {
       "epoch": 50.0,
-      "grad_norm": 2.4923367500305176,
+      "grad_norm": 2.4998741149902344,
       "learning_rate": 2.5e-05,
-      "loss": 0.0243,
+      "loss": 0.0251,
       "step": 5300
     },
     {
       "epoch": 50.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7866666666666666,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7659574468085107,
+      "eval_LOCATION_precision": 0.7468354430379747,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7105263157894737,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9084967320261438,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8910256410256411,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19272957742214203,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8328767123287671,
-      "eval_overall_precision": 0.793733681462141,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0636,
-      "eval_samples_per_second": 175.82,
-      "eval_steps_per_second": 2.821,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.16275979578495026,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8393351800554018,
+      "eval_overall_precision": 0.808,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4053,
+      "eval_samples_per_second": 461.392,
+      "eval_steps_per_second": 7.402,
       "step": 5300
     },
     {
       "epoch": 51.0,
-      "grad_norm": 0.5428743362426758,
+      "grad_norm": 1.0746209621429443,
       "learning_rate": 2.45e-05,
-      "loss": 0.0239,
+      "loss": 0.0251,
       "step": 5406
     },
     {
       "epoch": 51.0,
-      "eval_LOCATION_f1": 0.7483870967741935,
+      "eval_LOCATION_f1": 0.7763157894736843,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6904761904761905,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.7482993197278912,
+      "eval_LOCATION_precision": 0.7283950617283951,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7819548872180452,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6707317073170732,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_ORGANIZATION_precision": 0.7647058823529411,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7567567567567567,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6829268292682927,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8474576271186439,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8064516129032258,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.20626060664653778,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.8134228187919462,
-      "eval_overall_precision": 0.7613065326633166,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0562,
-      "eval_samples_per_second": 177.052,
-      "eval_steps_per_second": 2.84,
+      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.16527943313121796,
+      "eval_overall_accuracy": 0.9662948593598448,
+      "eval_overall_f1": 0.8444444444444444,
+      "eval_overall_precision": 0.8150134048257373,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4037,
+      "eval_samples_per_second": 463.197,
+      "eval_steps_per_second": 7.431,
       "step": 5406
     },
     {
       "epoch": 52.0,
-      "grad_norm": 0.7587451934814453,
+      "grad_norm": 1.7756356000900269,
       "learning_rate": 2.4e-05,
-      "loss": 0.0238,
+      "loss": 0.0245,
       "step": 5512
     },
     {
       "epoch": 52.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.7866666666666666,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7567567567567568,
+      "eval_LOCATION_precision": 0.7468354430379747,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.76,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6746987951807228,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8954248366013071,
+      "eval_ORGANIZATION_precision": 0.6705882352941176,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.9044585987261147,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8782051282051282,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7466666666666666,
+      "eval_PERSON_precision": 0.8658536585365854,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6666666666666666,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2040700912475586,
-      "eval_overall_accuracy": 0.9590203685741998,
-      "eval_overall_f1": 0.8276797829036635,
-      "eval_overall_precision": 0.782051282051282,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0567,
-      "eval_samples_per_second": 176.964,
-      "eval_steps_per_second": 2.839,
+      "eval_loss": 0.18333446979522705,
+      "eval_overall_accuracy": 0.9602327837051406,
+      "eval_overall_f1": 0.8306451612903226,
+      "eval_overall_precision": 0.7783375314861462,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.4031,
+      "eval_samples_per_second": 463.928,
+      "eval_steps_per_second": 7.443,
       "step": 5512
     },
     {
       "epoch": 53.0,
-      "grad_norm": 0.7874153852462769,
+      "grad_norm": 0.80136638879776,
       "learning_rate": 2.35e-05,
-      "loss": 0.0219,
+      "loss": 0.0222,
       "step": 5618
     },
     {
       "epoch": 53.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.759493670886076,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7534246575342466,
+      "eval_LOCATION_precision": 0.6896551724137931,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7346938775510204,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6790123456790124,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8954248366013071,
+      "eval_ORGANIZATION_precision": 0.6585365853658537,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9102564102564102,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8782051282051282,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8765432098765432,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8928571428571429,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_precision": 0.8620689655172413,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.19597817957401276,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8337874659400545,
-      "eval_overall_precision": 0.7906976744186046,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0499,
-      "eval_samples_per_second": 178.109,
-      "eval_steps_per_second": 2.857,
+      "eval_loss": 0.1887245774269104,
+      "eval_overall_accuracy": 0.9595053346265762,
+      "eval_overall_f1": 0.824631860776439,
+      "eval_overall_precision": 0.77,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4031,
+      "eval_samples_per_second": 463.882,
+      "eval_steps_per_second": 7.442,
       "step": 5618
     },
     {
       "epoch": 54.0,
-      "grad_norm": 1.4033691883087158,
+      "grad_norm": 1.39556086063385,
       "learning_rate": 2.3000000000000003e-05,
-      "loss": 0.0237,
+      "loss": 0.024,
       "step": 5724
     },
     {
       "epoch": 54.0,
-      "eval_LOCATION_f1": 0.8,
+      "eval_LOCATION_f1": 0.7919463087248323,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.759493670886076,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7651006711409396,
+      "eval_LOCATION_precision": 0.7564102564102564,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7682119205298013,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6785714285714286,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8932038834951457,
+      "eval_ORGANIZATION_precision": 0.6744186046511628,
+      "eval_ORGANIZATION_recall": 0.8923076923076924,
+      "eval_PERSON_f1": 0.9131832797427653,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8679245283018868,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8819875776397516,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19870668649673462,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8342391304347825,
-      "eval_overall_precision": 0.7892030848329049,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0458,
-      "eval_samples_per_second": 178.803,
-      "eval_steps_per_second": 2.869,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.17654088139533997,
+      "eval_overall_accuracy": 0.9633850630455868,
+      "eval_overall_f1": 0.8459459459459459,
+      "eval_overall_precision": 0.7964376590330788,
+      "eval_overall_recall": 0.9020172910662824,
+      "eval_runtime": 0.4024,
+      "eval_samples_per_second": 464.697,
+      "eval_steps_per_second": 7.455,
       "step": 5724
     },
     {
       "epoch": 55.0,
-      "grad_norm": 0.2151368260383606,
+      "grad_norm": 1.3348910808563232,
       "learning_rate": 2.25e-05,
-      "loss": 0.021,
+      "loss": 0.0238,
       "step": 5830
     },
     {
       "epoch": 55.0,
-      "eval_LOCATION_f1": 0.7692307692307694,
+      "eval_LOCATION_f1": 0.7866666666666666,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7058823529411765,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7468354430379748,
+      "eval_LOCATION_precision": 0.7468354430379747,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7755102040816327,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6344086021505376,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.6951219512195121,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.9190938511326862,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7012987012987013,
+      "eval_PERSON_precision": 0.8930817610062893,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6136363636363636,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.9259259259259259,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.9615384615384616,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.22194057703018188,
-      "eval_overall_accuracy": 0.9561105722599418,
-      "eval_overall_f1": 0.8126649076517151,
-      "eval_overall_precision": 0.7493917274939172,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0535,
-      "eval_samples_per_second": 177.497,
-      "eval_steps_per_second": 2.848,
+      "eval_loss": 0.1749105155467987,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8485675306957708,
+      "eval_overall_precision": 0.805699481865285,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.4051,
+      "eval_samples_per_second": 461.622,
+      "eval_steps_per_second": 7.406,
       "step": 5830
     },
     {
       "epoch": 56.0,
-      "grad_norm": 1.3246495723724365,
+      "grad_norm": 0.7062710523605347,
       "learning_rate": 2.2000000000000003e-05,
       "loss": 0.021,
       "step": 5936
     },
     {
       "epoch": 56.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7483870967741935,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7199999999999999,
+      "eval_LOCATION_precision": 0.6904761904761905,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7397260273972603,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6352941176470588,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_PERSON_f1": 0.9096774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_PERSON_precision": 0.88125,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8620689655172413,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8333333333333334,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.20850205421447754,
-      "eval_overall_accuracy": 0.9573229873908826,
-      "eval_overall_f1": 0.8091397849462366,
-      "eval_overall_precision": 0.7581863979848866,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.0545,
-      "eval_samples_per_second": 177.338,
-      "eval_steps_per_second": 2.845,
+      "eval_loss": 0.17994269728660583,
+      "eval_overall_accuracy": 0.9602327837051406,
+      "eval_overall_f1": 0.8259109311740891,
+      "eval_overall_precision": 0.7766497461928934,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4031,
+      "eval_samples_per_second": 463.918,
+      "eval_steps_per_second": 7.443,
       "step": 5936
     },
     {
       "epoch": 57.0,
-      "grad_norm": 1.0762540102005005,
+      "grad_norm": 0.6525917053222656,
       "learning_rate": 2.15e-05,
-      "loss": 0.0203,
+      "loss": 0.0199,
       "step": 6042
     },
     {
       "epoch": 57.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7763157894736843,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.76056338028169,
+      "eval_LOCATION_precision": 0.7283950617283951,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7012987012987013,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8938906752411575,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9102564102564102,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8633540372670807,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8765432098765432,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.9090909090909091,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2009107619524002,
-      "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.8276797829036635,
-      "eval_overall_precision": 0.782051282051282,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0514,
-      "eval_samples_per_second": 177.857,
-      "eval_steps_per_second": 2.853,
+      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.19421829283237457,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8422496570644719,
+      "eval_overall_precision": 0.8036649214659686,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4031,
+      "eval_samples_per_second": 463.878,
+      "eval_steps_per_second": 7.442,
       "step": 6042
     },
     {
       "epoch": 58.0,
-      "grad_norm": 1.3385218381881714,
+      "grad_norm": 1.2534284591674805,
       "learning_rate": 2.1e-05,
-      "loss": 0.0196,
+      "loss": 0.0187,
       "step": 6148
     },
     {
       "epoch": 58.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.75,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7464788732394366,
+      "eval_LOCATION_precision": 0.7037037037037037,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6883116883116883,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.6923076923076923,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.1974872499704361,
-      "eval_overall_accuracy": 0.960717749757517,
-      "eval_overall_f1": 0.8190476190476191,
-      "eval_overall_precision": 0.7757731958762887,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.06,
-      "eval_samples_per_second": 176.42,
-      "eval_steps_per_second": 2.83,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.19052091240882874,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8264462809917356,
+      "eval_overall_precision": 0.7915567282321899,
+      "eval_overall_recall": 0.8645533141210374,
+      "eval_runtime": 0.4084,
+      "eval_samples_per_second": 457.931,
+      "eval_steps_per_second": 7.346,
       "step": 6148
     },
     {
       "epoch": 59.0,
-      "grad_norm": 0.45897576212882996,
+      "grad_norm": 0.5067676305770874,
       "learning_rate": 2.05e-05,
-      "loss": 0.0195,
+      "loss": 0.0199,
       "step": 6254
     },
     {
       "epoch": 59.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7581699346405228,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7777777777777778,
+      "eval_LOCATION_precision": 0.7073170731707317,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7088607594936709,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9019607843137256,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_precision": 0.8846153846153846,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8070175438596492,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.20014727115631104,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.830393487109905,
-      "eval_overall_precision": 0.7846153846153846,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0622,
-      "eval_samples_per_second": 176.051,
-      "eval_steps_per_second": 2.824,
+      "eval_TIME_precision": 0.7931034482758621,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.19398272037506104,
+      "eval_overall_accuracy": 0.9592628516003879,
+      "eval_overall_f1": 0.8246575342465754,
+      "eval_overall_precision": 0.7859007832898173,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.404,
+      "eval_samples_per_second": 462.903,
+      "eval_steps_per_second": 7.426,
       "step": 6254
     },
     {
       "epoch": 60.0,
-      "grad_norm": 0.07298379391431808,
+      "grad_norm": 0.6053163409233093,
       "learning_rate": 2e-05,
-      "loss": 0.0188,
+      "loss": 0.0183,
       "step": 6360
     },
     {
       "epoch": 60.0,
-      "eval_LOCATION_f1": 0.8,
+      "eval_LOCATION_f1": 0.778523489932886,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.759493670886076,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7801418439716312,
+      "eval_LOCATION_precision": 0.7435897435897436,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7417218543046358,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7236842105263158,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8954248366013071,
+      "eval_ORGANIZATION_precision": 0.6511627906976745,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8782051282051282,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.6923076923076923,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.19487623870372772,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8363136176066025,
-      "eval_overall_precision": 0.8,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0468,
-      "eval_samples_per_second": 178.634,
-      "eval_steps_per_second": 2.866,
+      "eval_loss": 0.1951519399881363,
+      "eval_overall_accuracy": 0.9599903006789525,
+      "eval_overall_f1": 0.8276797829036635,
+      "eval_overall_precision": 0.782051282051282,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4039,
+      "eval_samples_per_second": 463.014,
+      "eval_steps_per_second": 7.428,
       "step": 6360
     },
     {
       "epoch": 61.0,
-      "grad_norm": 3.493285655975342,
+      "grad_norm": 0.19502782821655273,
       "learning_rate": 1.9500000000000003e-05,
-      "loss": 0.0197,
+      "loss": 0.0178,
       "step": 6466
     },
     {
       "epoch": 61.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.7638888888888888,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7619047619047619,
+      "eval_LOCATION_precision": 0.7534246575342466,
+      "eval_LOCATION_recall": 0.7746478873239436,
+      "eval_ORGANIZATION_f1": 0.7724137931034483,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6829268292682927,
+      "eval_ORGANIZATION_precision": 0.7,
       "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.887459807073955,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8571428571428571,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8450704225352113,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7894736842105263,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.20809414982795715,
-      "eval_overall_accuracy": 0.9595053346265762,
-      "eval_overall_f1": 0.827027027027027,
-      "eval_overall_precision": 0.7786259541984732,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.04,
-      "eval_samples_per_second": 179.808,
-      "eval_steps_per_second": 2.885,
+      "eval_loss": 0.190224289894104,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8386206896551723,
+      "eval_overall_precision": 0.8042328042328042,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4062,
+      "eval_samples_per_second": 460.355,
+      "eval_steps_per_second": 7.385,
       "step": 6466
     },
     {
       "epoch": 62.0,
-      "grad_norm": 0.8092681169509888,
+      "grad_norm": 0.6164102554321289,
       "learning_rate": 1.9e-05,
-      "loss": 0.0173,
+      "loss": 0.0196,
       "step": 6572
     },
     {
       "epoch": 62.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.778523489932886,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7567567567567568,
+      "eval_LOCATION_precision": 0.7435897435897436,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6746987951807228,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8214285714285714,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.20864509046077728,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8308525033829499,
-      "eval_overall_precision": 0.7831632653061225,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0523,
-      "eval_samples_per_second": 177.712,
-      "eval_steps_per_second": 2.851,
+      "eval_TIME_precision": 0.8214285714285714,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.18320535123348236,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8269230769230769,
+      "eval_overall_precision": 0.7900262467191601,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4026,
+      "eval_samples_per_second": 464.456,
+      "eval_steps_per_second": 7.451,
       "step": 6572
     },
     {
       "epoch": 63.0,
-      "grad_norm": 2.9163479804992676,
+      "grad_norm": 1.624141812324524,
       "learning_rate": 1.85e-05,
-      "loss": 0.0169,
+      "loss": 0.0182,
       "step": 6678
     },
     {
       "epoch": 63.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.7651006711409397,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.75,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7466666666666668,
+      "eval_LOCATION_precision": 0.7307692307692307,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7619047619047619,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6588235294117647,
+      "eval_ORGANIZATION_precision": 0.6829268292682927,
       "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8918032786885245,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8774193548387097,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
-      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_QUANTITY_precision": 0.6923076923076923,
+      "eval_QUANTITY_recall": 0.8181818181818182,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.20584411919116974,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8299319727891157,
-      "eval_overall_precision": 0.7860824742268041,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0618,
-      "eval_samples_per_second": 176.115,
-      "eval_steps_per_second": 2.825,
+      "eval_loss": 0.187962144613266,
+      "eval_overall_accuracy": 0.9616876818622696,
+      "eval_overall_f1": 0.8224043715846995,
+      "eval_overall_precision": 0.7818181818181819,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4039,
+      "eval_samples_per_second": 462.956,
+      "eval_steps_per_second": 7.427,
       "step": 6678
     },
     {
       "epoch": 64.0,
-      "grad_norm": 1.6419179439544678,
+      "grad_norm": 1.540419340133667,
       "learning_rate": 1.8e-05,
-      "loss": 0.0173,
+      "loss": 0.0181,
       "step": 6784
     },
     {
       "epoch": 64.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.7682119205298014,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.75,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.8,
+      "eval_LOCATION_precision": 0.725,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7586206896551724,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7714285714285715,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8852459016393444,
+      "eval_ORGANIZATION_precision": 0.6875,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8709677419354839,
-      "eval_PERSON_recall": 0.9,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
       "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.20184890925884247,
-      "eval_overall_accuracy": 0.9631425800193987,
-      "eval_overall_f1": 0.8372739916550765,
-      "eval_overall_precision": 0.8091397849462365,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.1318,
-      "eval_samples_per_second": 165.226,
-      "eval_steps_per_second": 2.651,
+      "eval_loss": 0.19288307428359985,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8233695652173914,
+      "eval_overall_precision": 0.7789203084832905,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4048,
+      "eval_samples_per_second": 461.933,
+      "eval_steps_per_second": 7.411,
       "step": 6784
     },
     {
       "epoch": 65.0,
-      "grad_norm": 0.9246335029602051,
+      "grad_norm": 0.8278260231018066,
       "learning_rate": 1.75e-05,
-      "loss": 0.0149,
+      "loss": 0.0187,
       "step": 6890
     },
     {
       "epoch": 65.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.7763157894736843,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.757142857142857,
+      "eval_LOCATION_precision": 0.7283950617283951,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7066666666666667,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_PERSON_f1": 0.8990228013029316,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8789808917197452,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_QUANTITY_precision": 0.6923076923076923,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_precision": 0.8928571428571429,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.20779463648796082,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8278688524590164,
-      "eval_overall_precision": 0.787012987012987,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0499,
-      "eval_samples_per_second": 178.113,
-      "eval_steps_per_second": 2.857,
+      "eval_loss": 0.19136402010917664,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8331034482758621,
+      "eval_overall_precision": 0.798941798941799,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.4041,
+      "eval_samples_per_second": 462.729,
+      "eval_steps_per_second": 7.423,
       "step": 6890
     },
     {
       "epoch": 66.0,
-      "grad_norm": 0.29005441069602966,
+      "grad_norm": 1.0269757509231567,
       "learning_rate": 1.7000000000000003e-05,
-      "loss": 0.0155,
+      "loss": 0.016,
       "step": 6996
     },
     {
       "epoch": 66.0,
-      "eval_LOCATION_f1": 0.7692307692307694,
+      "eval_LOCATION_f1": 0.7712418300653594,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7058823529411765,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7567567567567568,
+      "eval_LOCATION_precision": 0.7195121951219512,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6746987951807228,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22765351831912994,
-      "eval_overall_accuracy": 0.9592628516003879,
-      "eval_overall_f1": 0.8259109311740891,
-      "eval_overall_precision": 0.7766497461928934,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0539,
-      "eval_samples_per_second": 177.432,
-      "eval_steps_per_second": 2.847,
+      "eval_loss": 0.20218557119369507,
+      "eval_overall_accuracy": 0.9602327837051406,
+      "eval_overall_f1": 0.8244897959183674,
+      "eval_overall_precision": 0.7809278350515464,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4034,
+      "eval_samples_per_second": 463.539,
+      "eval_steps_per_second": 7.436,
       "step": 6996
     },
     {
       "epoch": 67.0,
-      "grad_norm": 0.15708033740520477,
+      "grad_norm": 0.6365354061126709,
       "learning_rate": 1.65e-05,
-      "loss": 0.0166,
+      "loss": 0.0153,
       "step": 7102
     },
     {
       "epoch": 67.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_PERSON_f1": 0.9096774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7567567567567567,
+      "eval_PERSON_precision": 0.88125,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6829268292682927,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21828927099704742,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8260869565217391,
-      "eval_overall_precision": 0.781491002570694,
+      "eval_loss": 0.19219978153705597,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.8317373461012312,
+      "eval_overall_precision": 0.7916666666666666,
       "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0465,
-      "eval_samples_per_second": 178.691,
-      "eval_steps_per_second": 2.867,
+      "eval_runtime": 0.4029,
+      "eval_samples_per_second": 464.137,
+      "eval_steps_per_second": 7.446,
       "step": 7102
     },
     {
       "epoch": 68.0,
-      "grad_norm": 0.745664656162262,
+      "grad_norm": 3.532707691192627,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.0158,
+      "loss": 0.0165,
       "step": 7208
     },
     {
       "epoch": 68.0,
-      "eval_LOCATION_f1": 0.7564102564102564,
+      "eval_LOCATION_f1": 0.76,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6941176470588235,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7534246575342466,
+      "eval_LOCATION_precision": 0.7215189873417721,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7567567567567568,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6790123456790124,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.896774193548387,
+      "eval_ORGANIZATION_precision": 0.6746987951807228,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_precision": 0.8853503184713376,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7466666666666666,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6666666666666666,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2362554669380188,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.8225806451612904,
-      "eval_overall_precision": 0.7707808564231738,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0542,
-      "eval_samples_per_second": 177.388,
-      "eval_steps_per_second": 2.846,
+      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.20774675905704498,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8256130790190734,
+      "eval_overall_precision": 0.7829457364341085,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4026,
+      "eval_samples_per_second": 464.494,
+      "eval_steps_per_second": 7.452,
       "step": 7208
     },
     {
       "epoch": 69.0,
-      "grad_norm": 0.1767091453075409,
+      "grad_norm": 1.3800381422042847,
       "learning_rate": 1.55e-05,
-      "loss": 0.0155,
+      "loss": 0.0159,
       "step": 7314
     },
     {
       "epoch": 69.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7534246575342466,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6790123456790124,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.821917808219178,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7945205479452054,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21981391310691833,
-      "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.8281461434370772,
-      "eval_overall_precision": 0.7806122448979592,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0623,
-      "eval_samples_per_second": 176.036,
-      "eval_steps_per_second": 2.824,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.20177708566188812,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8321964529331515,
+      "eval_overall_precision": 0.7901554404145078,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4055,
+      "eval_samples_per_second": 461.121,
+      "eval_steps_per_second": 7.398,
       "step": 7314
     },
     {
       "epoch": 70.0,
-      "grad_norm": 0.2924017906188965,
+      "grad_norm": 0.4369199573993683,
       "learning_rate": 1.5e-05,
-      "loss": 0.0155,
+      "loss": 0.0151,
       "step": 7420
     },
     {
       "epoch": 70.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7402597402597403,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7368421052631579,
+      "eval_LOCATION_precision": 0.6867469879518072,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6436781609195402,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9084967320261438,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_precision": 0.8910256410256411,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7567567567567567,
+      "eval_QUANTITY_f1": 0.7397260273972603,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6829268292682927,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8727272727272727,
+      "eval_QUANTITY_precision": 0.675,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2236405611038208,
+      "eval_loss": 0.21934795379638672,
       "eval_overall_accuracy": 0.9587778855480117,
-      "eval_overall_f1": 0.8274932614555257,
-      "eval_overall_precision": 0.7772151898734178,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0549,
-      "eval_samples_per_second": 177.26,
-      "eval_steps_per_second": 2.844,
+      "eval_overall_f1": 0.8212824010914053,
+      "eval_overall_precision": 0.7797927461139896,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4045,
+      "eval_samples_per_second": 462.289,
+      "eval_steps_per_second": 7.416,
       "step": 7420
     },
     {
       "epoch": 71.0,
-      "grad_norm": 0.8644899129867554,
+      "grad_norm": 0.6564136743545532,
       "learning_rate": 1.45e-05,
-      "loss": 0.0158,
+      "loss": 0.0149,
       "step": 7526
     },
     {
       "epoch": 71.0,
-      "eval_LOCATION_f1": 0.7564102564102564,
+      "eval_LOCATION_f1": 0.7763157894736843,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6941176470588235,
+      "eval_LOCATION_precision": 0.7283950617283951,
       "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7096774193548387,
+      "eval_ORGANIZATION_f1": 0.7671232876712328,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6111111111111112,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9025974025974027,
+      "eval_ORGANIZATION_precision": 0.691358024691358,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.879746835443038,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23344790935516357,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.8181818181818181,
-      "eval_overall_precision": 0.7630922693266833,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.1196,
-      "eval_samples_per_second": 167.021,
-      "eval_steps_per_second": 2.679,
+      "eval_loss": 0.21174688637256622,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8353741496598639,
+      "eval_overall_precision": 0.7912371134020618,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4036,
+      "eval_samples_per_second": 463.368,
+      "eval_steps_per_second": 7.434,
       "step": 7526
     },
     {
       "epoch": 72.0,
-      "grad_norm": 0.4155713617801666,
+      "grad_norm": 1.1040197610855103,
       "learning_rate": 1.4000000000000001e-05,
-      "loss": 0.0155,
+      "loss": 0.0152,
       "step": 7632
     },
     {
       "epoch": 72.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.7733333333333333,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_LOCATION_precision": 0.7341772151898734,
+      "eval_LOCATION_recall": 0.8169014084507042,
       "eval_ORGANIZATION_f1": 0.7746478873239436,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.7142857142857143,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_PERSON_f1": 0.9019607843137256,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8846153846153846,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8363636363636364,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21206453442573547,
-      "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.8278688524590164,
-      "eval_overall_precision": 0.787012987012987,
+      "eval_TIME_precision": 0.8518518518518519,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.19948169589042664,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8370165745856354,
+      "eval_overall_precision": 0.8037135278514589,
       "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0476,
-      "eval_samples_per_second": 178.503,
-      "eval_steps_per_second": 2.864,
+      "eval_runtime": 0.4055,
+      "eval_samples_per_second": 461.114,
+      "eval_steps_per_second": 7.398,
       "step": 7632
     },
     {
       "epoch": 73.0,
-      "grad_norm": 0.16996408998966217,
+      "grad_norm": 0.12047336250543594,
       "learning_rate": 1.3500000000000001e-05,
-      "loss": 0.0136,
+      "loss": 0.0149,
       "step": 7738
     },
     {
       "epoch": 73.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.7612903225806451,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_LOCATION_precision": 0.7023809523809523,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7619047619047619,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8938906752411575,
+      "eval_ORGANIZATION_precision": 0.6829268292682927,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8633540372670807,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7297297297297297,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.6585365853658537,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23162654042243958,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8238482384823849,
-      "eval_overall_precision": 0.7774936061381074,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0665,
-      "eval_samples_per_second": 175.338,
-      "eval_steps_per_second": 2.813,
+      "eval_loss": 0.2207670509815216,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8324324324324324,
+      "eval_overall_precision": 0.7837150127226463,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.4058,
+      "eval_samples_per_second": 460.775,
+      "eval_steps_per_second": 7.392,
       "step": 7738
     },
     {
       "epoch": 74.0,
-      "grad_norm": 1.7051116228103638,
+      "grad_norm": 1.3887410163879395,
       "learning_rate": 1.3000000000000001e-05,
-      "loss": 0.0134,
+      "loss": 0.0141,
       "step": 7844
     },
     {
       "epoch": 74.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.7567567567567568,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7857142857142856,
+      "eval_LOCATION_precision": 0.7272727272727273,
+      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_ORGANIZATION_f1": 0.7769784172661871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7333333333333333,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.7297297297297297,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22262556850910187,
-      "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.8379120879120878,
-      "eval_overall_precision": 0.800524934383202,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0491,
-      "eval_samples_per_second": 178.241,
-      "eval_steps_per_second": 2.859,
+      "eval_loss": 0.2141115963459015,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8370165745856354,
+      "eval_overall_precision": 0.8037135278514589,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4026,
+      "eval_samples_per_second": 464.486,
+      "eval_steps_per_second": 7.452,
       "step": 7844
     },
     {
       "epoch": 75.0,
-      "grad_norm": 0.22455532848834991,
+      "grad_norm": 0.22892561554908752,
       "learning_rate": 1.25e-05,
-      "loss": 0.0139,
+      "loss": 0.0142,
       "step": 7950
     },
     {
       "epoch": 75.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.7662337662337662,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
+      "eval_LOCATION_precision": 0.7108433734939759,
       "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7887323943661972,
+      "eval_ORGANIZATION_f1": 0.7222222222222222,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7272727272727273,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.6582278481012658,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21403230726718903,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.20973996818065643,
       "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.8356164383561643,
-      "eval_overall_precision": 0.7963446475195822,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0631,
-      "eval_samples_per_second": 175.908,
-      "eval_steps_per_second": 2.822,
+      "eval_overall_f1": 0.8283378746594005,
+      "eval_overall_precision": 0.7855297157622739,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4054,
+      "eval_samples_per_second": 461.245,
+      "eval_steps_per_second": 7.4,
       "step": 7950
     },
     {
       "epoch": 76.0,
-      "grad_norm": 0.007820017635822296,
+      "grad_norm": 0.28838738799095154,
       "learning_rate": 1.2e-05,
-      "loss": 0.0142,
+      "loss": 0.0132,
       "step": 8056
     },
     {
       "epoch": 76.0,
-      "eval_LOCATION_f1": 0.7692307692307694,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7058823529411765,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7746478873239436,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7517730496453899,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6973684210526315,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9096774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.88125,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22380225360393524,
-      "eval_overall_accuracy": 0.9612027158098934,
-      "eval_overall_f1": 0.8315217391304348,
-      "eval_overall_precision": 0.7866323907455013,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0493,
-      "eval_samples_per_second": 178.206,
-      "eval_steps_per_second": 2.859,
+      "eval_loss": 0.2148653268814087,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8333333333333333,
+      "eval_overall_precision": 0.7922077922077922,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.405,
+      "eval_samples_per_second": 461.757,
+      "eval_steps_per_second": 7.408,
       "step": 8056
     },
     {
       "epoch": 77.0,
-      "grad_norm": 0.41897082328796387,
+      "grad_norm": 0.2209300547838211,
       "learning_rate": 1.1500000000000002e-05,
-      "loss": 0.014,
+      "loss": 0.0132,
       "step": 8162
     },
     {
       "epoch": 77.0,
-      "eval_LOCATION_f1": 0.782051282051282,
+      "eval_LOCATION_f1": 0.7483870967741935,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7176470588235294,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.757142857142857,
+      "eval_LOCATION_precision": 0.6904761904761905,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7659574468085107,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7066666666666667,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.896774193548387,
+      "eval_ORGANIZATION_precision": 0.7105263157894737,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.86875,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8169014084507042,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7500000000000001,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7631578947368421,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.6923076923076923,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.8928571428571429,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21939164400100708,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.834924965893588,
-      "eval_overall_precision": 0.7927461139896373,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0552,
-      "eval_samples_per_second": 177.225,
-      "eval_steps_per_second": 2.843,
+      "eval_TIME_precision": 0.8928571428571429,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.21584409475326538,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.8321964529331515,
+      "eval_overall_precision": 0.7901554404145078,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4041,
+      "eval_samples_per_second": 462.714,
+      "eval_steps_per_second": 7.423,
       "step": 8162
     },
     {
       "epoch": 78.0,
-      "grad_norm": 0.47026127576828003,
+      "grad_norm": 0.5461943745613098,
       "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.0126,
+      "loss": 0.0141,
       "step": 8268
     },
     {
       "epoch": 78.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.75,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_PERSON_f1": 0.9019607843137256,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_precision": 0.8846153846153846,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_QUANTITY_f1": 0.7671232876712328,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_precision": 0.7,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22609065473079681,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8310626702997276,
-      "eval_overall_precision": 0.7881136950904393,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.052,
-      "eval_samples_per_second": 177.752,
-      "eval_steps_per_second": 2.852,
+      "eval_loss": 0.20878075063228607,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.8257887517146777,
+      "eval_overall_precision": 0.7879581151832461,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.4053,
+      "eval_samples_per_second": 461.363,
+      "eval_steps_per_second": 7.402,
       "step": 8268
     },
     {
       "epoch": 79.0,
-      "grad_norm": 0.9084616899490356,
+      "grad_norm": 0.6024172902107239,
       "learning_rate": 1.05e-05,
-      "loss": 0.0125,
+      "loss": 0.0117,
       "step": 8374
     },
     {
       "epoch": 79.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7586206896551724,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6875,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9061488673139159,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_precision": 0.8860759493670886,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
-      "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_QUANTITY_number": 33,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.232786163687706,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8281461434370772,
-      "eval_overall_precision": 0.7806122448979592,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0606,
-      "eval_samples_per_second": 176.309,
-      "eval_steps_per_second": 2.828,
+      "eval_loss": 0.20918363332748413,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8358620689655172,
+      "eval_overall_precision": 0.8015873015873016,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4026,
+      "eval_samples_per_second": 464.472,
+      "eval_steps_per_second": 7.451,
       "step": 8374
     },
     {
       "epoch": 80.0,
-      "grad_norm": 0.3104639947414398,
+      "grad_norm": 1.4614003896713257,
       "learning_rate": 1e-05,
-      "loss": 0.0123,
+      "loss": 0.0127,
       "step": 8480
     },
     {
       "epoch": 80.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7417218543046357,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7297297297297298,
+      "eval_LOCATION_precision": 0.7,
+      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6506024096385542,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9061488673139159,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_precision": 0.8860759493670886,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8888888888888888,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23563723266124725,
-      "eval_overall_accuracy": 0.9597478176527643,
-      "eval_overall_f1": 0.8247978436657682,
-      "eval_overall_precision": 0.7746835443037975,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0481,
-      "eval_samples_per_second": 178.419,
-      "eval_steps_per_second": 2.862,
+      "eval_loss": 0.2205442190170288,
+      "eval_overall_accuracy": 0.9587778855480117,
+      "eval_overall_f1": 0.8273972602739728,
+      "eval_overall_precision": 0.7885117493472585,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.4027,
+      "eval_samples_per_second": 464.311,
+      "eval_steps_per_second": 7.449,
       "step": 8480
     },
     {
       "epoch": 81.0,
-      "grad_norm": 0.09613505750894547,
+      "grad_norm": 1.915273666381836,
       "learning_rate": 9.5e-06,
-      "loss": 0.0121,
+      "loss": 0.0134,
       "step": 8586
     },
     {
       "epoch": 81.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.76,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7857142857142856,
+      "eval_LOCATION_precision": 0.7215189873417721,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7333333333333333,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8852459016393444,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8709677419354839,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21699251234531403,
-      "eval_overall_accuracy": 0.9633850630455868,
-      "eval_overall_f1": 0.8335625859697388,
-      "eval_overall_precision": 0.7973684210526316,
+      "eval_loss": 0.20929260551929474,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8370165745856354,
+      "eval_overall_precision": 0.8037135278514589,
       "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0455,
-      "eval_samples_per_second": 178.863,
-      "eval_steps_per_second": 2.869,
+      "eval_runtime": 0.4028,
+      "eval_samples_per_second": 464.264,
+      "eval_steps_per_second": 7.448,
       "step": 8586
     },
     {
       "epoch": 82.0,
-      "grad_norm": 1.2361814975738525,
+      "grad_norm": 1.0925920009613037,
       "learning_rate": 9e-06,
-      "loss": 0.0124,
+      "loss": 0.0138,
       "step": 8692
     },
     {
       "epoch": 82.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7682119205298014,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7397260273972603,
+      "eval_LOCATION_precision": 0.725,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7659574468085107,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_precision": 0.7105263157894737,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9032258064516129,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.875,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23771944642066956,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8259109311740891,
-      "eval_overall_precision": 0.7766497461928934,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0536,
-      "eval_samples_per_second": 177.487,
-      "eval_steps_per_second": 2.847,
+      "eval_loss": 0.20777982473373413,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.8351648351648351,
+      "eval_overall_precision": 0.7979002624671916,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.404,
+      "eval_samples_per_second": 462.894,
+      "eval_steps_per_second": 7.426,
       "step": 8692
     },
     {
       "epoch": 83.0,
-      "grad_norm": 0.15847894549369812,
+      "grad_norm": 1.2260353565216064,
       "learning_rate": 8.500000000000002e-06,
-      "loss": 0.0123,
+      "loss": 0.0129,
       "step": 8798
     },
     {
       "epoch": 83.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7417218543046357,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7397260273972603,
+      "eval_LOCATION_precision": 0.7,
+      "eval_LOCATION_recall": 0.7887323943661971,
+      "eval_ORGANIZATION_f1": 0.7692307692307693,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9032258064516129,
+      "eval_ORGANIZATION_precision": 0.7051282051282052,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.875,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2313675880432129,
-      "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.827027027027027,
-      "eval_overall_precision": 0.7786259541984732,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0661,
-      "eval_samples_per_second": 175.406,
-      "eval_steps_per_second": 2.814,
+      "eval_loss": 0.21702027320861816,
+      "eval_overall_accuracy": 0.9599903006789525,
+      "eval_overall_f1": 0.825136612021858,
+      "eval_overall_precision": 0.7844155844155845,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.4043,
+      "eval_samples_per_second": 462.499,
+      "eval_steps_per_second": 7.42,
       "step": 8798
     },
     {
       "epoch": 84.0,
-      "grad_norm": 0.36980700492858887,
+      "grad_norm": 0.29825663566589355,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.0121,
+      "loss": 0.0119,
       "step": 8904
     },
     {
       "epoch": 84.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7712418300653594,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7746478873239436,
+      "eval_LOCATION_precision": 0.7195121951219512,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
+      "eval_PERSON_precision": 0.8734177215189873,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21659883856773376,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.8387978142076504,
-      "eval_overall_precision": 0.7974025974025974,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.055,
-      "eval_samples_per_second": 177.255,
-      "eval_steps_per_second": 2.844,
+      "eval_loss": 0.2102896124124527,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.8301369863013699,
+      "eval_overall_precision": 0.7911227154046997,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.4038,
+      "eval_samples_per_second": 463.06,
+      "eval_steps_per_second": 7.429,
       "step": 8904
     },
     {
       "epoch": 85.0,
-      "grad_norm": 0.6816823482513428,
+      "grad_norm": 2.5998857021331787,
       "learning_rate": 7.5e-06,
-      "loss": 0.0122,
+      "loss": 0.0117,
       "step": 9010
     },
     {
       "epoch": 85.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7516778523489933,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7755102040816327,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9090909090909091,
+      "eval_ORGANIZATION_precision": 0.6951219512195121,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8860759493670886,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7945205479452054,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8070175438596492,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.24010856449604034,
-      "eval_overall_accuracy": 0.9599903006789525,
-      "eval_overall_f1": 0.82907133243607,
-      "eval_overall_precision": 0.7777777777777778,
+      "eval_TIME_precision": 0.7931034482758621,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.22094765305519104,
+      "eval_overall_accuracy": 0.9597478176527643,
+      "eval_overall_f1": 0.8346883468834688,
+      "eval_overall_precision": 0.7877237851662404,
       "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0528,
-      "eval_samples_per_second": 177.616,
-      "eval_steps_per_second": 2.849,
+      "eval_runtime": 0.4049,
+      "eval_samples_per_second": 461.865,
+      "eval_steps_per_second": 7.41,
       "step": 9010
     },
     {
       "epoch": 86.0,
-      "grad_norm": 1.010087490081787,
+      "grad_norm": 1.0339630842208862,
       "learning_rate": 7.000000000000001e-06,
-      "loss": 0.012,
+      "loss": 0.0129,
       "step": 9116
     },
     {
       "epoch": 86.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7746478873239436,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7659574468085107,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.894736842105263,
+      "eval_ORGANIZATION_precision": 0.7105263157894737,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8831168831168831,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2242480218410492,
-      "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.8340192043895748,
-      "eval_overall_precision": 0.7958115183246073,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0597,
-      "eval_samples_per_second": 176.469,
-      "eval_steps_per_second": 2.831,
+      "eval_loss": 0.2100267857313156,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8372093023255814,
+      "eval_overall_precision": 0.796875,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4047,
+      "eval_samples_per_second": 462.12,
+      "eval_steps_per_second": 7.414,
       "step": 9116
     },
     {
       "epoch": 87.0,
-      "grad_norm": 0.7689788937568665,
+      "grad_norm": 0.0695308968424797,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.0112,
+      "loss": 0.0107,
       "step": 9222
     },
     {
       "epoch": 87.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7432432432432433,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7692307692307693,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6626506024096386,
+      "eval_ORGANIZATION_precision": 0.7051282051282052,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7777777777777778,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.717948717948718,
-      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2339707463979721,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.827027027027027,
-      "eval_overall_precision": 0.7786259541984732,
+      "eval_loss": 0.21508151292800903,
+      "eval_overall_accuracy": 0.9616876818622696,
+      "eval_overall_f1": 0.8372093023255814,
+      "eval_overall_precision": 0.796875,
       "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 1.0569,
-      "eval_samples_per_second": 176.936,
-      "eval_steps_per_second": 2.839,
+      "eval_runtime": 0.407,
+      "eval_samples_per_second": 459.502,
+      "eval_steps_per_second": 7.372,
       "step": 9222
     },
     {
       "epoch": 88.0,
-      "grad_norm": 0.36982235312461853,
+      "grad_norm": 1.1862696409225464,
       "learning_rate": 6e-06,
-      "loss": 0.0109,
+      "loss": 0.0121,
       "step": 9328
     },
     {
       "epoch": 88.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7733333333333333,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7341772151898734,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7659574468085107,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.7105263157894737,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8990228013029316,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8789808917197452,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22819143533706665,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8217687074829931,
-      "eval_overall_precision": 0.7783505154639175,
+      "eval_loss": 0.21262787282466888,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8319559228650137,
+      "eval_overall_precision": 0.7968337730870713,
       "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0643,
-      "eval_samples_per_second": 175.699,
-      "eval_steps_per_second": 2.819,
+      "eval_runtime": 0.4061,
+      "eval_samples_per_second": 460.423,
+      "eval_steps_per_second": 7.386,
       "step": 9328
     },
     {
       "epoch": 89.0,
-      "grad_norm": 0.2916944622993469,
+      "grad_norm": 1.132206678390503,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.0105,
+      "loss": 0.0104,
       "step": 9434
     },
     {
       "epoch": 89.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7733333333333333,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7464788732394366,
+      "eval_LOCATION_precision": 0.7341772151898734,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7857142857142856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6883116883116883,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
       "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.8860759493670886,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
+      "eval_QUANTITY_precision": 0.7368421052631579,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2318750023841858,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8288043478260869,
-      "eval_overall_precision": 0.7840616966580977,
+      "eval_loss": 0.21018344163894653,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8402203856749312,
+      "eval_overall_precision": 0.8047493403693932,
       "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0577,
-      "eval_samples_per_second": 176.795,
-      "eval_steps_per_second": 2.836,
+      "eval_runtime": 0.4076,
+      "eval_samples_per_second": 458.732,
+      "eval_steps_per_second": 7.359,
       "step": 9434
     },
     {
       "epoch": 90.0,
-      "grad_norm": 0.37730202078819275,
+      "grad_norm": 0.5784242749214172,
       "learning_rate": 5e-06,
-      "loss": 0.0098,
+      "loss": 0.0114,
       "step": 9540
     },
     {
       "epoch": 90.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7682119205298014,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7346938775510204,
+      "eval_LOCATION_precision": 0.725,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7826086956521738,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6585365853658537,
+      "eval_ORGANIZATION_precision": 0.7397260273972602,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
-      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23395000398159027,
-      "eval_overall_accuracy": 0.960717749757517,
-      "eval_overall_f1": 0.8276797829036635,
-      "eval_overall_precision": 0.782051282051282,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0495,
-      "eval_samples_per_second": 178.178,
-      "eval_steps_per_second": 2.858,
+      "eval_loss": 0.21030160784721375,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.837465564738292,
+      "eval_overall_precision": 0.8021108179419525,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.4106,
+      "eval_samples_per_second": 455.383,
+      "eval_steps_per_second": 7.306,
       "step": 9540
     },
     {
       "epoch": 91.0,
-      "grad_norm": 0.6458882689476013,
+      "grad_norm": 0.3709010183811188,
       "learning_rate": 4.5e-06,
-      "loss": 0.0109,
+      "loss": 0.0105,
       "step": 9646
     },
     {
       "epoch": 91.0,
-      "eval_LOCATION_f1": 0.7843137254901961,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7317073170731707,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7586206896551724,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7769784172661871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6875,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.7297297297297297,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7671232876712328,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7,
-      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.22991229593753815,
-      "eval_overall_accuracy": 0.9612027158098934,
-      "eval_overall_f1": 0.8310626702997276,
-      "eval_overall_precision": 0.7881136950904393,
+      "eval_loss": 0.2143879383802414,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.8367626886145404,
+      "eval_overall_precision": 0.7984293193717278,
       "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 1.0583,
-      "eval_samples_per_second": 176.704,
-      "eval_steps_per_second": 2.835,
+      "eval_runtime": 0.4069,
+      "eval_samples_per_second": 459.523,
+      "eval_steps_per_second": 7.372,
       "step": 9646
     },
     {
       "epoch": 92.0,
-      "grad_norm": 0.31122303009033203,
+      "grad_norm": 0.19695152342319489,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.0114,
+      "loss": 0.011,
       "step": 9752
     },
     {
       "epoch": 92.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.75,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7037037037037037,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7692307692307693,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.7051282051282052,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23025508224964142,
-      "eval_overall_accuracy": 0.9597478176527643,
-      "eval_overall_f1": 0.8190476190476191,
-      "eval_overall_precision": 0.7757731958762887,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.046,
-      "eval_samples_per_second": 178.77,
-      "eval_steps_per_second": 2.868,
+      "eval_loss": 0.22280381619930267,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4088,
+      "eval_samples_per_second": 457.432,
+      "eval_steps_per_second": 7.338,
       "step": 9752
     },
     {
       "epoch": 93.0,
-      "grad_norm": 1.2301055192947388,
+      "grad_norm": 0.30122533440589905,
       "learning_rate": 3.5000000000000004e-06,
-      "loss": 0.0104,
+      "loss": 0.0107,
       "step": 9858
     },
     {
       "epoch": 93.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7346938775510204,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6585365853658537,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23245875537395477,
-      "eval_overall_accuracy": 0.9599903006789525,
-      "eval_overall_f1": 0.8189189189189189,
-      "eval_overall_precision": 0.7709923664122137,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0547,
-      "eval_samples_per_second": 177.306,
-      "eval_steps_per_second": 2.844,
+      "eval_loss": 0.2211771309375763,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4094,
+      "eval_samples_per_second": 456.814,
+      "eval_steps_per_second": 7.329,
       "step": 9858
     },
     {
       "epoch": 94.0,
-      "grad_norm": 1.0476171970367432,
+      "grad_norm": 0.1013573631644249,
       "learning_rate": 3e-06,
-      "loss": 0.0099,
+      "loss": 0.009,
       "step": 9964
     },
     {
       "epoch": 94.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7586206896551724,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6875,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23353543877601624,
-      "eval_overall_accuracy": 0.9595053346265762,
-      "eval_overall_f1": 0.8179347826086958,
-      "eval_overall_precision": 0.7737789203084833,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.0587,
-      "eval_samples_per_second": 176.634,
-      "eval_steps_per_second": 2.834,
+      "eval_loss": 0.2232343554496765,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8353741496598639,
+      "eval_overall_precision": 0.7912371134020618,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4075,
+      "eval_samples_per_second": 458.923,
+      "eval_steps_per_second": 7.362,
       "step": 9964
     },
     {
       "epoch": 95.0,
-      "grad_norm": 0.6297835111618042,
+      "grad_norm": 1.0435041189193726,
       "learning_rate": 2.5e-06,
-      "loss": 0.0115,
+      "loss": 0.0109,
       "step": 10070
     },
     {
       "epoch": 95.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7450980392156863,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7310344827586207,
+      "eval_LOCATION_precision": 0.6951219512195121,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6625,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.24176815152168274,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.22744058072566986,
       "eval_overall_accuracy": 0.9592628516003879,
-      "eval_overall_f1": 0.8173207036535859,
-      "eval_overall_precision": 0.7704081632653061,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.0575,
-      "eval_samples_per_second": 176.827,
-      "eval_steps_per_second": 2.837,
+      "eval_overall_f1": 0.8306010928961749,
+      "eval_overall_precision": 0.7896103896103897,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.407,
+      "eval_samples_per_second": 459.405,
+      "eval_steps_per_second": 7.37,
       "step": 10070
     },
     {
       "epoch": 96.0,
-      "grad_norm": 0.7284921407699585,
+      "grad_norm": 4.114624500274658,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.009,
+      "loss": 0.0098,
       "step": 10176
     },
     {
       "epoch": 96.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7631578947368421,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_LOCATION_precision": 0.7160493827160493,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7692307692307693,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.7051282051282052,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.23633427917957306,
-      "eval_overall_accuracy": 0.9592628516003879,
-      "eval_overall_f1": 0.8179347826086958,
-      "eval_overall_precision": 0.7737789203084833,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.0622,
-      "eval_samples_per_second": 176.042,
-      "eval_steps_per_second": 2.824,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.22330667078495026,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8387978142076504,
+      "eval_overall_precision": 0.7974025974025974,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.4058,
+      "eval_samples_per_second": 460.857,
+      "eval_steps_per_second": 7.393,
       "step": 10176
     },
     {
       "epoch": 97.0,
-      "grad_norm": 0.8745085000991821,
+      "grad_norm": 0.8480021953582764,
       "learning_rate": 1.5e-06,
-      "loss": 0.0104,
+      "loss": 0.0103,
       "step": 10282
     },
     {
       "epoch": 97.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7346938775510204,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.763888888888889,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6585365853658537,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2431822121143341,
-      "eval_overall_accuracy": 0.9597478176527643,
-      "eval_overall_f1": 0.8194070080862534,
-      "eval_overall_precision": 0.769620253164557,
+      "eval_loss": 0.22035245597362518,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8283378746594005,
+      "eval_overall_precision": 0.7855297157622739,
       "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 1.0389,
-      "eval_samples_per_second": 180.0,
-      "eval_steps_per_second": 2.888,
+      "eval_runtime": 0.4072,
+      "eval_samples_per_second": 459.182,
+      "eval_steps_per_second": 7.367,
       "step": 10282
     },
     {
       "epoch": 98.0,
-      "grad_norm": 0.7874897122383118,
+      "grad_norm": 0.7559491395950317,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 0.0104,
+      "loss": 0.0102,
       "step": 10388
     },
     {
       "epoch": 98.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7682119205298014,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_LOCATION_precision": 0.725,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.763888888888889,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2386394888162613,
-      "eval_overall_accuracy": 0.9590203685741998,
-      "eval_overall_f1": 0.8179347826086958,
-      "eval_overall_precision": 0.7737789203084833,
-      "eval_overall_recall": 0.8674351585014409,
-      "eval_runtime": 1.0472,
-      "eval_samples_per_second": 178.565,
-      "eval_steps_per_second": 2.865,
+      "eval_loss": 0.2218533307313919,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4081,
+      "eval_samples_per_second": 458.203,
+      "eval_steps_per_second": 7.351,
       "step": 10388
     },
     {
       "epoch": 99.0,
-      "grad_norm": 0.21786460280418396,
+      "grad_norm": 0.24367158114910126,
       "learning_rate": 5.000000000000001e-07,
-      "loss": 0.0106,
+      "loss": 0.0094,
       "step": 10494
     },
     {
       "epoch": 99.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.7549668874172185,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7397260273972603,
+      "eval_LOCATION_precision": 0.7125,
+      "eval_LOCATION_recall": 0.8028169014084507,
+      "eval_ORGANIZATION_f1": 0.763888888888889,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.24056798219680786,
-      "eval_overall_accuracy": 0.9597478176527643,
-      "eval_overall_f1": 0.8200270635994588,
-      "eval_overall_precision": 0.7729591836734694,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0478,
-      "eval_samples_per_second": 178.469,
-      "eval_steps_per_second": 2.863,
+      "eval_loss": 0.22336292266845703,
+      "eval_overall_accuracy": 0.9609602327837051,
+      "eval_overall_f1": 0.8321964529331515,
+      "eval_overall_precision": 0.7901554404145078,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.4081,
+      "eval_samples_per_second": 458.197,
+      "eval_steps_per_second": 7.351,
       "step": 10494
     },
     {
       "epoch": 100.0,
-      "grad_norm": 0.5074381828308105,
+      "grad_norm": 0.0760447159409523,
       "learning_rate": 0.0,
-      "loss": 0.0093,
+      "loss": 0.0098,
       "step": 10600
     },
     {
       "epoch": 100.0,
-      "eval_LOCATION_f1": 0.7741935483870968,
+      "eval_LOCATION_f1": 0.7682119205298014,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7142857142857143,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7397260273972603,
+      "eval_LOCATION_precision": 0.725,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.763888888888889,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7397260273972603,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.675,
-      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.24040785431861877,
-      "eval_overall_accuracy": 0.9592628516003879,
-      "eval_overall_f1": 0.8173207036535859,
-      "eval_overall_precision": 0.7704081632653061,
-      "eval_overall_recall": 0.8703170028818443,
-      "eval_runtime": 1.05,
-      "eval_samples_per_second": 178.101,
-      "eval_steps_per_second": 2.857,
+      "eval_loss": 0.22373317182064056,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.4055,
+      "eval_samples_per_second": 461.127,
+      "eval_steps_per_second": 7.398,
       "step": 10600
     },
     {
       "epoch": 100.0,
       "step": 10600,
       "total_flos": 5189413292285520.0,
-      "train_loss": 0.05698325344414081,
-      "train_runtime": 1582.8095,
-      "train_samples_per_second": 106.646,
-      "train_steps_per_second": 6.697
+      "train_loss": 0.055730826843459653,
+      "train_runtime": 962.8691,
+      "train_samples_per_second": 175.309,
+      "train_steps_per_second": 11.009
     }
   ],
   "logging_steps": 500,