diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -10,282 +10,282 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 3.6827080249786377,
+      "grad_norm": 4.570859432220459,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 0.3501,
+      "loss": 0.341,
       "step": 106
     },
     {
       "epoch": 1.0,
-      "eval_LOCATION_f1": 0.7483870967741935,
+      "eval_LOCATION_f1": 0.8,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.6904761904761905,
-      "eval_LOCATION_recall": 0.8169014084507042,
-      "eval_ORGANIZATION_f1": 0.6709677419354838,
+      "eval_LOCATION_precision": 0.7380952380952381,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.6973684210526315,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5777777777777777,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8679245283018867,
+      "eval_ORGANIZATION_precision": 0.6091954022988506,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8509316770186336,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8214285714285714,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7123287671232875,
+      "eval_PERSON_precision": 0.7965116279069767,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7341772151898734,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.65,
-      "eval_QUANTITY_recall": 0.7878787878787878,
-      "eval_TIME_f1": 0.896551724137931,
+      "eval_QUANTITY_precision": 0.6304347826086957,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.7692307692307693,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8666666666666667,
-      "eval_TIME_recall": 0.9285714285714286,
-      "eval_loss": 0.13337989151477814,
-      "eval_overall_accuracy": 0.953443258971872,
-      "eval_overall_f1": 0.7905138339920948,
-      "eval_overall_precision": 0.7281553398058253,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 0.9448,
-      "eval_samples_per_second": 197.919,
-      "eval_steps_per_second": 3.175,
+      "eval_TIME_precision": 0.6756756756756757,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.1448812186717987,
+      "eval_overall_accuracy": 0.9522308438409312,
+      "eval_overall_f1": 0.7917205692108668,
+      "eval_overall_precision": 0.7183098591549296,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3665,
+      "eval_samples_per_second": 510.216,
+      "eval_steps_per_second": 8.185,
       "step": 106
     },
     {
       "epoch": 2.0,
-      "grad_norm": 2.115112781524658,
+      "grad_norm": 2.158717393875122,
       "learning_rate": 4.9e-05,
-      "loss": 0.1148,
+      "loss": 0.125,
       "step": 212
     },
     {
       "epoch": 2.0,
-      "eval_LOCATION_f1": 0.7612903225806451,
+      "eval_LOCATION_f1": 0.7875000000000001,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7023809523809523,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.697986577181208,
+      "eval_LOCATION_precision": 0.7078651685393258,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.6707317073170732,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6190476190476191,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_ORGANIZATION_precision": 0.5555555555555556,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8974358974358974,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.6329113924050633,
+      "eval_PERSON_precision": 0.8641975308641975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.6506024096385542,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.5434782608695652,
-      "eval_QUANTITY_recall": 0.7575757575757576,
-      "eval_TIME_f1": 0.7692307692307693,
+      "eval_QUANTITY_precision": 0.54,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "eval_TIME_f1": 0.7187499999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6756756756756757,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.1362091302871704,
-      "eval_overall_accuracy": 0.9558680892337537,
-      "eval_overall_f1": 0.7926023778071334,
-      "eval_overall_precision": 0.7317073170731707,
-      "eval_overall_recall": 0.8645533141210374,
-      "eval_runtime": 0.9986,
-      "eval_samples_per_second": 187.27,
-      "eval_steps_per_second": 3.004,
+      "eval_TIME_precision": 0.6388888888888888,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.15502017736434937,
+      "eval_overall_accuracy": 0.9500484966052376,
+      "eval_overall_f1": 0.7867177522349936,
+      "eval_overall_precision": 0.7064220183486238,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3859,
+      "eval_samples_per_second": 484.571,
+      "eval_steps_per_second": 7.774,
       "step": 212
     },
     {
       "epoch": 3.0,
-      "grad_norm": 2.7631375789642334,
+      "grad_norm": 3.599668025970459,
       "learning_rate": 4.85e-05,
-      "loss": 0.0732,
+      "loss": 0.0885,
       "step": 318
     },
     {
       "epoch": 3.0,
-      "eval_LOCATION_f1": 0.7662337662337662,
+      "eval_LOCATION_f1": 0.725,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7108433734939759,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7702702702702702,
+      "eval_LOCATION_precision": 0.651685393258427,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7837837837837838,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6867469879518072,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.6987951807228916,
+      "eval_ORGANIZATION_recall": 0.8923076923076924,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8115942028985507,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.806451612903226,
+      "eval_QUANTITY_precision": 0.7777777777777778,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7352941176470589,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.14412528276443481,
-      "eval_overall_accuracy": 0.9573229873908826,
-      "eval_overall_f1": 0.8221024258760108,
-      "eval_overall_precision": 0.7721518987341772,
+      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.140259250998497,
+      "eval_overall_accuracy": 0.9561105722599418,
+      "eval_overall_f1": 0.8254397834912043,
+      "eval_overall_precision": 0.7780612244897959,
       "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 0.9223,
-      "eval_samples_per_second": 202.756,
-      "eval_steps_per_second": 3.253,
+      "eval_runtime": 0.3708,
+      "eval_samples_per_second": 504.338,
+      "eval_steps_per_second": 8.091,
       "step": 318
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.3901716470718384,
+      "grad_norm": 0.3577202558517456,
       "learning_rate": 4.8e-05,
-      "loss": 0.0453,
+      "loss": 0.0577,
       "step": 424
     },
     {
       "epoch": 4.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8053691275167785,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7123287671232877,
+      "eval_LOCATION_precision": 0.7692307692307693,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7333333333333334,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6419753086419753,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_ORGANIZATION_precision": 0.6470588235294118,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8974358974358974,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7837837837837839,
+      "eval_PERSON_precision": 0.8641975308641975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8115942028985507,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7073170731707317,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8214285714285714,
+      "eval_QUANTITY_precision": 0.7777777777777778,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8214285714285714,
-      "eval_TIME_recall": 0.8214285714285714,
-      "eval_loss": 0.1702796369791031,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.8260869565217391,
-      "eval_overall_precision": 0.781491002570694,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 0.94,
-      "eval_samples_per_second": 198.941,
-      "eval_steps_per_second": 3.192,
+      "eval_TIME_precision": 0.8888888888888888,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.159939706325531,
+      "eval_overall_accuracy": 0.9604752667313288,
+      "eval_overall_f1": 0.8353741496598639,
+      "eval_overall_precision": 0.7912371134020618,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3999,
+      "eval_samples_per_second": 467.655,
+      "eval_steps_per_second": 7.502,
       "step": 424
     },
     {
       "epoch": 5.0,
-      "grad_norm": 0.532292902469635,
+      "grad_norm": 2.663273572921753,
       "learning_rate": 4.75e-05,
-      "loss": 0.0312,
+      "loss": 0.0416,
       "step": 530
     },
     {
       "epoch": 5.0,
-      "eval_LOCATION_f1": 0.8051948051948051,
+      "eval_LOCATION_f1": 0.7894736842105262,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7469879518072289,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7832167832167831,
+      "eval_LOCATION_precision": 0.7407407407407407,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7333333333333334,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.717948717948718,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.6470588235294118,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7692307692307693,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.17514730989933014,
-      "eval_overall_accuracy": 0.9612027158098934,
-      "eval_overall_f1": 0.8324324324324324,
-      "eval_overall_precision": 0.7837150127226463,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9647,
-      "eval_samples_per_second": 193.847,
-      "eval_steps_per_second": 3.11,
+      "eval_TIME_precision": 0.6756756756756757,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.19336794316768646,
+      "eval_overall_accuracy": 0.9546556741028128,
+      "eval_overall_f1": 0.8192771084337349,
+      "eval_overall_precision": 0.765,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3772,
+      "eval_samples_per_second": 495.755,
+      "eval_steps_per_second": 7.953,
       "step": 530
     },
     {
       "epoch": 6.0,
-      "grad_norm": 5.477240085601807,
+      "grad_norm": 1.2095777988433838,
       "learning_rate": 4.7e-05,
-      "loss": 0.0236,
+      "loss": 0.0279,
       "step": 636
     },
     {
       "epoch": 6.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.8025477707006369,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7482993197278912,
+      "eval_LOCATION_precision": 0.7325581395348837,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7142857142857142,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6707317073170732,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.90032154340836,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.8698412698412699,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8695652173913043,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7837837837837839,
+      "eval_PERSON_precision": 0.8303030303030303,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7073170731707317,
+      "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.75,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6666666666666666,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.20889906585216522,
-      "eval_overall_accuracy": 0.9570805043646945,
-      "eval_overall_f1": 0.8197596795727637,
-      "eval_overall_precision": 0.763681592039801,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9433,
-      "eval_samples_per_second": 198.231,
-      "eval_steps_per_second": 3.18,
+      "eval_loss": 0.2035682201385498,
+      "eval_overall_accuracy": 0.9556256062075654,
+      "eval_overall_f1": 0.814516129032258,
+      "eval_overall_precision": 0.7632241813602015,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.3709,
+      "eval_samples_per_second": 504.187,
+      "eval_steps_per_second": 8.089,
       "step": 636
     },
     {
       "epoch": 7.0,
-      "grad_norm": 2.7495665550231934,
+      "grad_norm": 2.685880661010742,
       "learning_rate": 4.6500000000000005e-05,
-      "loss": 0.017,
+      "loss": 0.0217,
       "step": 742
     },
     {
       "epoch": 7.0,
-      "eval_LOCATION_f1": 0.8275862068965518,
+      "eval_LOCATION_f1": 0.7692307692307694,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8108108108108109,
+      "eval_LOCATION_precision": 0.7058823529411765,
       "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.763888888888889,
+      "eval_ORGANIZATION_f1": 0.7517730496453899,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6962025316455697,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6973684210526315,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8838709677419355,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_precision": 0.85625,
       "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8108108108108109,
+      "eval_QUANTITY_f1": 0.7605633802816901,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7317073170731707,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7105263157894737,
+      "eval_QUANTITY_recall": 0.8181818181818182,
+      "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": 0.21119879186153412,
-      "eval_overall_accuracy": 0.957807953443259,
-      "eval_overall_f1": 0.8406593406593407,
-      "eval_overall_precision": 0.8031496062992126,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9486,
-      "eval_samples_per_second": 197.142,
-      "eval_steps_per_second": 3.163,
+      "eval_loss": 0.2385089248418808,
+      "eval_overall_accuracy": 0.9536857419980601,
+      "eval_overall_f1": 0.8190476190476191,
+      "eval_overall_precision": 0.7757731958762887,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.3706,
+      "eval_samples_per_second": 504.653,
+      "eval_steps_per_second": 8.096,
       "step": 742
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.43028321862220764,
+      "grad_norm": 1.1274532079696655,
       "learning_rate": 4.600000000000001e-05,
-      "loss": 0.0133,
+      "loss": 0.0156,
       "step": 848
     },
     {
@@ -294,895 +294,895 @@
       "eval_LOCATION_number": 71,
       "eval_LOCATION_precision": 0.775,
       "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7862068965517242,
+      "eval_ORGANIZATION_f1": 0.6906474820143885,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7125,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_ORGANIZATION_precision": 0.6486486486486487,
+      "eval_ORGANIZATION_recall": 0.7384615384615385,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8108108108108109,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7317073170731707,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7666666666666666,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.21715472638607025,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8466757123473542,
-      "eval_overall_precision": 0.8,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9553,
-      "eval_samples_per_second": 195.744,
-      "eval_steps_per_second": 3.14,
+      "eval_TIME_precision": 0.71875,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.24100376665592194,
+      "eval_overall_accuracy": 0.9580504364694471,
+      "eval_overall_f1": 0.8246575342465754,
+      "eval_overall_precision": 0.7859007832898173,
+      "eval_overall_recall": 0.8674351585014409,
+      "eval_runtime": 0.3806,
+      "eval_samples_per_second": 491.299,
+      "eval_steps_per_second": 7.882,
       "step": 848
     },
     {
       "epoch": 9.0,
-      "grad_norm": 0.03873327746987343,
+      "grad_norm": 0.09234097599983215,
       "learning_rate": 4.55e-05,
-      "loss": 0.0096,
+      "loss": 0.0115,
       "step": 954
     },
     {
       "epoch": 9.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.802721088435374,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7482993197278912,
+      "eval_LOCATION_precision": 0.7763157894736842,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6707317073170732,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8874172185430463,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7887323943661972,
+      "eval_PERSON_precision": 0.881578947368421,
+      "eval_PERSON_recall": 0.8933333333333333,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7368421052631579,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8474576271186439,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8064516129032258,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.23949624598026276,
-      "eval_overall_accuracy": 0.9585354025218235,
-      "eval_overall_f1": 0.8365122615803814,
-      "eval_overall_precision": 0.7932816537467701,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0163,
-      "eval_samples_per_second": 183.996,
-      "eval_steps_per_second": 2.952,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.2652629017829895,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8363636363636364,
+      "eval_overall_precision": 0.8125,
+      "eval_overall_recall": 0.861671469740634,
+      "eval_runtime": 0.3686,
+      "eval_samples_per_second": 507.285,
+      "eval_steps_per_second": 8.138,
       "step": 954
     },
     {
       "epoch": 10.0,
-      "grad_norm": 1.984230875968933,
+      "grad_norm": 0.10166473686695099,
       "learning_rate": 4.5e-05,
-      "loss": 0.0077,
+      "loss": 0.0115,
       "step": 1060
     },
     {
       "epoch": 10.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.7945205479452055,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7297297297297298,
+      "eval_LOCATION_precision": 0.7733333333333333,
+      "eval_LOCATION_recall": 0.8169014084507042,
+      "eval_ORGANIZATION_f1": 0.7887323943661972,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6506024096385542,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8996763754045306,
+      "eval_ORGANIZATION_precision": 0.7272727272727273,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_precision": 0.879746835443038,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_QUANTITY_f1": 0.8450704225352113,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.806451612903226,
+      "eval_QUANTITY_precision": 0.7894736842105263,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7352941176470589,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.26252666115760803,
-      "eval_overall_accuracy": 0.9568380213385063,
-      "eval_overall_f1": 0.8268456375838926,
-      "eval_overall_precision": 0.7738693467336684,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.952,
-      "eval_samples_per_second": 196.421,
-      "eval_steps_per_second": 3.151,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.23218387365341187,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8480662983425414,
+      "eval_overall_precision": 0.8143236074270557,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.381,
+      "eval_samples_per_second": 490.801,
+      "eval_steps_per_second": 7.874,
       "step": 1060
     },
     {
       "epoch": 11.0,
-      "grad_norm": 0.09802401065826416,
+      "grad_norm": 3.547823905944824,
       "learning_rate": 4.4500000000000004e-05,
-      "loss": 0.0059,
+      "loss": 0.0073,
       "step": 1166
     },
     {
       "epoch": 11.0,
-      "eval_LOCATION_f1": 0.8400000000000001,
+      "eval_LOCATION_f1": 0.8053691275167785,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7974683544303798,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7659574468085107,
+      "eval_LOCATION_precision": 0.7692307692307693,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.802919708029197,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7105263157894737,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8823529411764707,
+      "eval_ORGANIZATION_precision": 0.7638888888888888,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8653846153846154,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8169014084507042,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8732394366197183,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7631578947368421,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8474576271186439,
+      "eval_QUANTITY_precision": 0.8157894736842105,
+      "eval_QUANTITY_recall": 0.9393939393939394,
+      "eval_TIME_f1": 0.912280701754386,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8064516129032258,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.235382542014122,
-      "eval_overall_accuracy": 0.9648399612027158,
-      "eval_overall_f1": 0.8418156808803301,
-      "eval_overall_precision": 0.8052631578947368,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9545,
-      "eval_samples_per_second": 195.917,
-      "eval_steps_per_second": 3.143,
+      "eval_TIME_precision": 0.896551724137931,
+      "eval_TIME_recall": 0.9285714285714286,
+      "eval_loss": 0.23406679928302765,
+      "eval_overall_accuracy": 0.9636275460717749,
+      "eval_overall_f1": 0.8630705394190872,
+      "eval_overall_precision": 0.8297872340425532,
+      "eval_overall_recall": 0.899135446685879,
+      "eval_runtime": 0.3709,
+      "eval_samples_per_second": 504.149,
+      "eval_steps_per_second": 8.088,
       "step": 1166
     },
     {
       "epoch": 12.0,
-      "grad_norm": 0.4382927417755127,
+      "grad_norm": 4.043017864227295,
       "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.0078,
+      "loss": 0.0061,
       "step": 1272
     },
     {
       "epoch": 12.0,
-      "eval_LOCATION_f1": 0.8251748251748252,
+      "eval_LOCATION_f1": 0.8053691275167785,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8194444444444444,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.8201438848920863,
+      "eval_LOCATION_precision": 0.7692307692307693,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7801418439716312,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7702702702702703,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.7236842105263158,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9038461538461539,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8115942028985507,
+      "eval_PERSON_precision": 0.8703703703703703,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7777777777777778,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8813559322033899,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.21610549092292786,
-      "eval_overall_accuracy": 0.9675072744907857,
-      "eval_overall_f1": 0.8575418994413407,
-      "eval_overall_precision": 0.8319783197831978,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9504,
-      "eval_samples_per_second": 196.751,
-      "eval_steps_per_second": 3.156,
+      "eval_TIME_precision": 0.8387096774193549,
+      "eval_TIME_recall": 0.9285714285714286,
+      "eval_loss": 0.26555678248405457,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8485675306957708,
+      "eval_overall_precision": 0.805699481865285,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3727,
+      "eval_samples_per_second": 501.749,
+      "eval_steps_per_second": 8.049,
       "step": 1272
     },
     {
       "epoch": 13.0,
-      "grad_norm": 2.4125523567199707,
+      "grad_norm": 1.4539355039596558,
       "learning_rate": 4.35e-05,
-      "loss": 0.0044,
+      "loss": 0.0076,
       "step": 1378
     },
     {
       "epoch": 13.0,
-      "eval_LOCATION_f1": 0.8181818181818181,
+      "eval_LOCATION_f1": 0.8082191780821917,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7590361445783133,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7552447552447553,
+      "eval_LOCATION_precision": 0.7866666666666666,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6923076923076923,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.821917808219178,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8813559322033899,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8387096774193549,
-      "eval_TIME_recall": 0.9285714285714286,
-      "eval_loss": 0.26299452781677246,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8469945355191256,
-      "eval_overall_precision": 0.8051948051948052,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9454,
-      "eval_samples_per_second": 197.792,
-      "eval_steps_per_second": 3.173,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.2568497955799103,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8464730290456433,
+      "eval_overall_precision": 0.8138297872340425,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3728,
+      "eval_samples_per_second": 501.558,
+      "eval_steps_per_second": 8.046,
       "step": 1378
     },
     {
       "epoch": 14.0,
-      "grad_norm": 0.6766614317893982,
+      "grad_norm": 0.1553322672843933,
       "learning_rate": 4.3e-05,
-      "loss": 0.0047,
+      "loss": 0.008,
       "step": 1484
     },
     {
       "epoch": 14.0,
-      "eval_LOCATION_f1": 0.8493150684931506,
+      "eval_LOCATION_f1": 0.8,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8266666666666667,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7862068965517242,
+      "eval_LOCATION_precision": 0.759493670886076,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7755102040816327,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7125,
+      "eval_ORGANIZATION_precision": 0.6951219512195121,
       "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_PERSON_f1": 0.8938906752411575,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_PERSON_precision": 0.8633540372670807,
+      "eval_PERSON_recall": 0.9266666666666666,
       "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.9090909090909091,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.9259259259259259,
+      "eval_TIME_precision": 0.8064516129032258,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.25089046359062195,
-      "eval_overall_accuracy": 0.9638700290979632,
-      "eval_overall_f1": 0.8500687757909215,
-      "eval_overall_precision": 0.8131578947368421,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9494,
-      "eval_samples_per_second": 196.976,
-      "eval_steps_per_second": 3.16,
+      "eval_loss": 0.2785702347755432,
+      "eval_overall_accuracy": 0.9599903006789525,
+      "eval_overall_f1": 0.8389715832205683,
+      "eval_overall_precision": 0.7908163265306123,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3781,
+      "eval_samples_per_second": 494.547,
+      "eval_steps_per_second": 7.934,
       "step": 1484
     },
     {
       "epoch": 15.0,
-      "grad_norm": 2.052441358566284,
+      "grad_norm": 0.5594255924224854,
       "learning_rate": 4.25e-05,
-      "loss": 0.0041,
+      "loss": 0.0061,
       "step": 1590
     },
     {
       "epoch": 15.0,
-      "eval_LOCATION_f1": 0.8104575163398693,
+      "eval_LOCATION_f1": 0.7894736842105262,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7560975609756098,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7755102040816327,
+      "eval_LOCATION_precision": 0.7407407407407407,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7189542483660131,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6951219512195121,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8831168831168831,
+      "eval_ORGANIZATION_precision": 0.625,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8607594936708861,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.821917808219178,
+      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.7619047619047619,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8771929824561403,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6857142857142857,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2882152795791626,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8306451612903226,
-      "eval_overall_precision": 0.7783375314861462,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9611,
-      "eval_samples_per_second": 194.573,
-      "eval_steps_per_second": 3.121,
+      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.2651939392089844,
+      "eval_overall_accuracy": 0.9590203685741998,
+      "eval_overall_f1": 0.8243243243243243,
+      "eval_overall_precision": 0.7760814249363868,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.3698,
+      "eval_samples_per_second": 505.726,
+      "eval_steps_per_second": 8.113,
       "step": 1590
     },
     {
       "epoch": 16.0,
-      "grad_norm": 0.014275839552283287,
+      "grad_norm": 1.9039210081100464,
       "learning_rate": 4.2e-05,
-      "loss": 0.0044,
+      "loss": 0.0033,
       "step": 1696
     },
     {
       "epoch": 16.0,
-      "eval_LOCATION_f1": 0.8053691275167785,
+      "eval_LOCATION_f1": 0.8211920529801324,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.775,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7769784172661871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8910891089108911,
+      "eval_ORGANIZATION_precision": 0.7297297297297297,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.90032154340836,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8823529411764706,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8169014084507042,
+      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8450704225352113,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7631578947368421,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7741935483870968,
+      "eval_QUANTITY_precision": 0.7894736842105263,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7058823529411765,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.26748523116111755,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8372093023255814,
-      "eval_overall_precision": 0.796875,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9645,
-      "eval_samples_per_second": 193.888,
-      "eval_steps_per_second": 3.111,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.28800952434539795,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.8497267759562841,
+      "eval_overall_precision": 0.8077922077922078,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3789,
+      "eval_samples_per_second": 493.556,
+      "eval_steps_per_second": 7.918,
       "step": 1696
     },
     {
       "epoch": 17.0,
-      "grad_norm": 0.5238301157951355,
+      "grad_norm": 3.2790563106536865,
       "learning_rate": 4.15e-05,
-      "loss": 0.0032,
+      "loss": 0.004,
       "step": 1802
     },
     {
       "epoch": 17.0,
-      "eval_LOCATION_f1": 0.8299319727891157,
+      "eval_LOCATION_f1": 0.8400000000000001,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8026315789473685,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7746478873239436,
+      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.8057553956834532,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8896103896103896,
+      "eval_ORGANIZATION_precision": 0.7567567567567568,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9084967320261438,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8670886075949367,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8000000000000002,
+      "eval_PERSON_precision": 0.8910256410256411,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8115942028985507,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_precision": 0.7777777777777778,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7619047619047619,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6857142857142857,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.28771573305130005,
-      "eval_overall_accuracy": 0.9612027158098934,
-      "eval_overall_f1": 0.8356164383561643,
-      "eval_overall_precision": 0.7963446475195822,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 0.9538,
-      "eval_samples_per_second": 196.063,
-      "eval_steps_per_second": 3.145,
+      "eval_loss": 0.2645246684551239,
+      "eval_overall_accuracy": 0.9655674102812803,
+      "eval_overall_f1": 0.8563535911602209,
+      "eval_overall_precision": 0.8222811671087533,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3682,
+      "eval_samples_per_second": 507.831,
+      "eval_steps_per_second": 8.147,
       "step": 1802
     },
     {
       "epoch": 18.0,
-      "grad_norm": 0.005304540041834116,
+      "grad_norm": 0.05785554274916649,
       "learning_rate": 4.1e-05,
-      "loss": 0.0034,
+      "loss": 0.0036,
       "step": 1908
     },
     {
       "epoch": 18.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.8378378378378378,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7552447552447553,
+      "eval_LOCATION_precision": 0.8051948051948052,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7857142857142856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6923076923076923,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.894736842105263,
+      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9038461538461539,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8831168831168831,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.821917808219178,
+      "eval_PERSON_precision": 0.8703703703703703,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8928571428571429,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.793103448275862,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8928571428571429,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.28492802381515503,
-      "eval_overall_accuracy": 0.9614451988360815,
-      "eval_overall_f1": 0.8429752066115703,
-      "eval_overall_precision": 0.8073878627968337,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9617,
-      "eval_samples_per_second": 194.443,
-      "eval_steps_per_second": 3.119,
+      "eval_TIME_precision": 0.7666666666666667,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.30065634846687317,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8477366255144033,
+      "eval_overall_precision": 0.8089005235602095,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3713,
+      "eval_samples_per_second": 503.695,
+      "eval_steps_per_second": 8.081,
       "step": 1908
     },
     {
       "epoch": 19.0,
-      "grad_norm": 0.05048967897891998,
+      "grad_norm": 0.16370488703250885,
       "learning_rate": 4.05e-05,
-      "loss": 0.0051,
+      "loss": 0.0036,
       "step": 2014
     },
     {
       "epoch": 19.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.8211920529801324,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7832167832167831,
+      "eval_LOCATION_precision": 0.775,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.717948717948718,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8918032786885245,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8774193548387097,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2656005918979645,
-      "eval_overall_accuracy": 0.9643549951503395,
-      "eval_overall_f1": 0.8493150684931507,
-      "eval_overall_precision": 0.8093994778067886,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9518,
-      "eval_samples_per_second": 196.476,
-      "eval_steps_per_second": 3.152,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.2996464669704437,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8354080221300139,
+      "eval_overall_precision": 0.8031914893617021,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.3681,
+      "eval_samples_per_second": 508.021,
+      "eval_steps_per_second": 8.15,
       "step": 2014
     },
     {
       "epoch": 20.0,
-      "grad_norm": 0.005232055671513081,
+      "grad_norm": 0.06265392154455185,
       "learning_rate": 4e-05,
-      "loss": 0.0029,
+      "loss": 0.0063,
       "step": 2120
     },
     {
       "epoch": 20.0,
-      "eval_LOCATION_f1": 0.7948717948717949,
+      "eval_LOCATION_f1": 0.7922077922077921,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7294117647058823,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7482014388489208,
+      "eval_LOCATION_precision": 0.7349397590361446,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7027027027027027,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.9032258064516129,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8974358974358974,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_precision": 0.8641975308641975,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.823529411764706,
+      "eval_QUANTITY_f1": 0.8115942028985507,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8,
+      "eval_QUANTITY_precision": 0.7777777777777778,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.819672131147541,
+      "eval_TIME_f1": 0.8363636363636364,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7575757575757576,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2749578654766083,
-      "eval_overall_accuracy": 0.9638700290979632,
-      "eval_overall_f1": 0.8365122615803814,
-      "eval_overall_precision": 0.7932816537467701,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9399,
-      "eval_samples_per_second": 198.964,
-      "eval_steps_per_second": 3.192,
+      "eval_TIME_precision": 0.8518518518518519,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.28041526675224304,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8402203856749312,
+      "eval_overall_precision": 0.8047493403693932,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.3881,
+      "eval_samples_per_second": 481.834,
+      "eval_steps_per_second": 7.73,
       "step": 2120
     },
     {
       "epoch": 21.0,
-      "grad_norm": 20.893596649169922,
+      "grad_norm": 0.020481407642364502,
       "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.0051,
+      "loss": 0.0034,
       "step": 2226
     },
     {
       "epoch": 21.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.8187919463087248,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
+      "eval_LOCATION_precision": 0.782051282051282,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7464788732394366,
+      "eval_ORGANIZATION_f1": 0.7832167832167831,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6883116883116883,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.717948717948718,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8115942028985507,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7777777777777778,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_precision": 0.8064516129032258,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2634299099445343,
-      "eval_overall_accuracy": 0.9643549951503395,
-      "eval_overall_f1": 0.8429752066115703,
-      "eval_overall_precision": 0.8073878627968337,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9692,
-      "eval_samples_per_second": 192.946,
-      "eval_steps_per_second": 3.095,
+      "eval_loss": 0.2882232367992401,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8532235939643347,
+      "eval_overall_precision": 0.8141361256544503,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3668,
+      "eval_samples_per_second": 509.838,
+      "eval_steps_per_second": 8.179,
       "step": 2226
     },
     {
       "epoch": 22.0,
-      "grad_norm": 0.013275664299726486,
+      "grad_norm": 0.18097683787345886,
       "learning_rate": 3.9000000000000006e-05,
       "loss": 0.003,
       "step": 2332
     },
     {
       "epoch": 22.0,
-      "eval_LOCATION_f1": 0.8243243243243243,
+      "eval_LOCATION_f1": 0.8079470198675496,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7922077922077922,
+      "eval_LOCATION_precision": 0.7625,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7417218543046358,
+      "eval_ORGANIZATION_f1": 0.7862068965517242,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6511627906976745,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9061488673139159,
+      "eval_ORGANIZATION_precision": 0.7125,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_precision": 0.8974358974358975,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8333333333333333,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7692307692307693,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2914196252822876,
-      "eval_overall_accuracy": 0.9614451988360815,
-      "eval_overall_f1": 0.8455284552845529,
-      "eval_overall_precision": 0.7979539641943734,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9669,
-      "eval_samples_per_second": 193.405,
-      "eval_steps_per_second": 3.103,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.31117871403694153,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8508891928864569,
+      "eval_overall_precision": 0.8098958333333334,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3742,
+      "eval_samples_per_second": 499.76,
+      "eval_steps_per_second": 8.018,
       "step": 2332
     },
     {
       "epoch": 23.0,
-      "grad_norm": 0.05714813619852066,
+      "grad_norm": 21.788358688354492,
       "learning_rate": 3.85e-05,
-      "loss": 0.0035,
+      "loss": 0.004,
       "step": 2438
     },
     {
       "epoch": 23.0,
-      "eval_LOCATION_f1": 0.8435374149659864,
+      "eval_LOCATION_f1": 0.7712418300653594,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8157894736842105,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.76056338028169,
+      "eval_LOCATION_precision": 0.7195121951219512,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7462686567164178,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7012987012987013,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.7246376811594203,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
+      "eval_PERSON_precision": 0.879746835443038,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8115942028985507,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7777777777777778,
+      "eval_QUANTITY_precision": 0.7368421052631579,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_TIME_f1": 0.7796610169491526,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2690264582633972,
-      "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8543689320388349,
-      "eval_overall_precision": 0.8235294117647058,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.943,
-      "eval_samples_per_second": 198.31,
-      "eval_steps_per_second": 3.181,
+      "eval_TIME_precision": 0.7419354838709677,
+      "eval_TIME_recall": 0.8214285714285714,
+      "eval_loss": 0.2963677644729614,
+      "eval_overall_accuracy": 0.9616876818622696,
+      "eval_overall_f1": 0.8248275862068966,
+      "eval_overall_precision": 0.791005291005291,
+      "eval_overall_recall": 0.861671469740634,
+      "eval_runtime": 0.3729,
+      "eval_samples_per_second": 501.526,
+      "eval_steps_per_second": 8.046,
       "step": 2438
     },
     {
       "epoch": 24.0,
-      "grad_norm": 8.982757568359375,
+      "grad_norm": 0.08339240401983261,
       "learning_rate": 3.8e-05,
-      "loss": 0.0024,
+      "loss": 0.0037,
       "step": 2544
     },
     {
       "epoch": 24.0,
-      "eval_LOCATION_f1": 0.7924528301886792,
+      "eval_LOCATION_f1": 0.8079470198675496,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7159090909090909,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7549668874172185,
+      "eval_LOCATION_precision": 0.7625,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7162162162162162,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6627906976744186,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9108910891089109,
+      "eval_ORGANIZATION_precision": 0.6385542168674698,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.894736842105263,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9019607843137255,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8333333333333333,
+      "eval_PERSON_precision": 0.8831168831168831,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.8571428571428571,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7692307692307693,
+      "eval_QUANTITY_precision": 0.8108108108108109,
       "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.301530122756958,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8425302826379543,
-      "eval_overall_precision": 0.7904040404040404,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.9593,
-      "eval_samples_per_second": 194.925,
-      "eval_steps_per_second": 3.127,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.2963782250881195,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8306010928961749,
+      "eval_overall_precision": 0.7896103896103897,
+      "eval_overall_recall": 0.8760806916426513,
+      "eval_runtime": 0.374,
+      "eval_samples_per_second": 500.061,
+      "eval_steps_per_second": 8.022,
       "step": 2544
     },
     {
       "epoch": 25.0,
-      "grad_norm": 0.10265465825796127,
+      "grad_norm": 0.022947581484913826,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0032,
+      "loss": 0.0048,
       "step": 2650
     },
     {
       "epoch": 25.0,
-      "eval_LOCATION_f1": 0.8157894736842104,
+      "eval_LOCATION_f1": 0.8289473684210527,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7654320987654321,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7971014492753623,
+      "eval_LOCATION_precision": 0.7777777777777778,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7368421052631577,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7534246575342466,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9,
+      "eval_ORGANIZATION_precision": 0.7205882352941176,
+      "eval_ORGANIZATION_recall": 0.7538461538461538,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8656716417910447,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_precision": 0.8529411764705882,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2815525233745575,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.851182197496523,
-      "eval_overall_precision": 0.8225806451612904,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9507,
-      "eval_samples_per_second": 196.688,
-      "eval_steps_per_second": 3.155,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.26159676909446716,
+      "eval_overall_accuracy": 0.9660523763336566,
+      "eval_overall_f1": 0.853146853146853,
+      "eval_overall_precision": 0.8288043478260869,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.375,
+      "eval_samples_per_second": 498.713,
+      "eval_steps_per_second": 8.001,
       "step": 2650
     },
     {
       "epoch": 26.0,
-      "grad_norm": 0.0032238499261438847,
+      "grad_norm": 0.005224139895290136,
       "learning_rate": 3.7e-05,
-      "loss": 0.003,
+      "loss": 0.0041,
       "step": 2756
     },
     {
       "epoch": 26.0,
-      "eval_LOCATION_f1": 0.8344370860927152,
+      "eval_LOCATION_f1": 0.8181818181818181,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_precision": 0.7590361445783133,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7659574468085107,
+      "eval_ORGANIZATION_f1": 0.7375886524822695,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7105263157894737,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.6842105263157895,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8333333333333333,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8529411764705883,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7692307692307693,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_QUANTITY_precision": 0.8285714285714286,
+      "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": 0.28177666664123535,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.8508287292817679,
-      "eval_overall_precision": 0.8169761273209549,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9649,
-      "eval_samples_per_second": 193.801,
-      "eval_steps_per_second": 3.109,
+      "eval_loss": 0.2906554937362671,
+      "eval_overall_accuracy": 0.965082444228904,
+      "eval_overall_f1": 0.8500687757909215,
+      "eval_overall_precision": 0.8131578947368421,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3676,
+      "eval_samples_per_second": 508.66,
+      "eval_steps_per_second": 8.16,
       "step": 2756
     },
     {
       "epoch": 27.0,
-      "grad_norm": 0.03518635779619217,
+      "grad_norm": 0.013817857019603252,
       "learning_rate": 3.65e-05,
-      "loss": 0.0016,
+      "loss": 0.0038,
       "step": 2862
     },
     {
       "epoch": 27.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.7973856209150327,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
+      "eval_LOCATION_precision": 0.7439024390243902,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7777777777777778,
+      "eval_ORGANIZATION_f1": 0.7194244604316546,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7088607594936709,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.6756756756756757,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
       "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2842986285686493,
-      "eval_overall_accuracy": 0.9648399612027158,
-      "eval_overall_f1": 0.8528198074277854,
-      "eval_overall_precision": 0.8157894736842105,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 1.03,
-      "eval_samples_per_second": 181.562,
-      "eval_steps_per_second": 2.913,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.3264944851398468,
+      "eval_overall_accuracy": 0.9619301648884578,
+      "eval_overall_f1": 0.8367626886145404,
+      "eval_overall_precision": 0.7984293193717278,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.3664,
+      "eval_samples_per_second": 510.385,
+      "eval_steps_per_second": 8.188,
       "step": 2862
     },
     {
       "epoch": 28.0,
-      "grad_norm": 0.01968814805150032,
+      "grad_norm": 0.911362886428833,
       "learning_rate": 3.6e-05,
-      "loss": 0.0025,
+      "loss": 0.0046,
       "step": 2968
     },
     {
       "epoch": 28.0,
-      "eval_LOCATION_f1": 0.8366013071895425,
+      "eval_LOCATION_f1": 0.7763157894736843,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7804878048780488,
-      "eval_LOCATION_recall": 0.9014084507042254,
-      "eval_ORGANIZATION_f1": 0.8028169014084509,
+      "eval_LOCATION_precision": 0.7283950617283951,
+      "eval_LOCATION_recall": 0.8309859154929577,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7402597402597403,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8853503184713376,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7714285714285715,
+      "eval_PERSON_precision": 0.8475609756097561,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7297297297297297,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.8571428571428571,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3238835334777832,
-      "eval_overall_accuracy": 0.960717749757517,
-      "eval_overall_f1": 0.8512396694214875,
-      "eval_overall_precision": 0.8153034300791556,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9375,
-      "eval_samples_per_second": 199.463,
-      "eval_steps_per_second": 3.2,
+      "eval_loss": 0.30422839522361755,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8301369863013699,
+      "eval_overall_precision": 0.7911227154046997,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.368,
+      "eval_samples_per_second": 508.171,
+      "eval_steps_per_second": 8.152,
       "step": 2968
     },
     {
       "epoch": 29.0,
-      "grad_norm": 0.005619656760245562,
+      "grad_norm": 0.25442686676979065,
       "learning_rate": 3.55e-05,
-      "loss": 0.0026,
+      "loss": 0.003,
       "step": 3074
     },
     {
       "epoch": 29.0,
-      "eval_LOCATION_f1": 0.8421052631578947,
+      "eval_LOCATION_f1": 0.8079470198675496,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7901234567901234,
-      "eval_LOCATION_recall": 0.9014084507042254,
-      "eval_ORGANIZATION_f1": 0.8111888111888114,
+      "eval_LOCATION_precision": 0.7625,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7435897435897436,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.875,
+      "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": 0.8474576271186439,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8064516129032258,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.3090417683124542,
-      "eval_overall_accuracy": 0.965082444228904,
-      "eval_overall_f1": 0.8575342465753425,
-      "eval_overall_precision": 0.8172323759791122,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.9535,
-      "eval_samples_per_second": 196.122,
-      "eval_steps_per_second": 3.146,
+      "eval_TIME_precision": 0.8,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.32693979144096375,
+      "eval_overall_accuracy": 0.960717749757517,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3676,
+      "eval_samples_per_second": 508.67,
+      "eval_steps_per_second": 8.16,
       "step": 3074
     },
     {
       "epoch": 30.0,
-      "grad_norm": 0.0475270040333271,
+      "grad_norm": 0.3953845202922821,
       "learning_rate": 3.5e-05,
-      "loss": 0.0041,
+      "loss": 0.0031,
       "step": 3180
     },
     {
       "epoch": 30.0,
-      "eval_LOCATION_f1": 0.8211920529801324,
+      "eval_LOCATION_f1": 0.8051948051948051,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.775,
+      "eval_LOCATION_precision": 0.7469879518072289,
       "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7703703703703704,
+      "eval_ORGANIZATION_f1": 0.7164179104477613,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7428571428571429,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.6956521739130435,
+      "eval_ORGANIZATION_recall": 0.7384615384615385,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.30209869146347046,
-      "eval_overall_accuracy": 0.9626576139670223,
-      "eval_overall_f1": 0.8464730290456433,
-      "eval_overall_precision": 0.8138297872340425,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9526,
-      "eval_samples_per_second": 196.299,
-      "eval_steps_per_second": 3.149,
+      "eval_loss": 0.29468849301338196,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8347107438016529,
+      "eval_overall_precision": 0.7994722955145118,
+      "eval_overall_recall": 0.8731988472622478,
+      "eval_runtime": 0.368,
+      "eval_samples_per_second": 508.123,
+      "eval_steps_per_second": 8.152,
       "step": 3180
     },
     {
       "epoch": 31.0,
-      "grad_norm": 0.6186401844024658,
+      "grad_norm": 0.3862725496292114,
       "learning_rate": 3.45e-05,
-      "loss": 0.0054,
+      "loss": 0.0026,
       "step": 3286
     },
     {
@@ -1191,1067 +1191,1067 @@
       "eval_LOCATION_number": 71,
       "eval_LOCATION_precision": 0.7590361445783133,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.759124087591241,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7222222222222222,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.9025974025974027,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9114754098360656,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_precision": 0.896774193548387,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.75,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6666666666666666,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3126040995121002,
-      "eval_overall_accuracy": 0.9595053346265762,
-      "eval_overall_f1": 0.8353741496598639,
-      "eval_overall_precision": 0.7912371134020618,
+      "eval_loss": 0.2897959053516388,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8457300275482094,
+      "eval_overall_precision": 0.8100263852242744,
       "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9488,
-      "eval_samples_per_second": 197.092,
-      "eval_steps_per_second": 3.162,
+      "eval_runtime": 0.369,
+      "eval_samples_per_second": 506.776,
+      "eval_steps_per_second": 8.13,
       "step": 3286
     },
     {
       "epoch": 32.0,
-      "grad_norm": 0.001890498329885304,
+      "grad_norm": 0.0018400972476229072,
       "learning_rate": 3.4000000000000007e-05,
-      "loss": 0.002,
+      "loss": 0.0024,
       "step": 3392
     },
     {
       "epoch": 32.0,
-      "eval_LOCATION_f1": 0.8243243243243243,
+      "eval_LOCATION_f1": 0.8157894736842104,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7922077922077922,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7737226277372263,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7297297297297298,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7361111111111112,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_ORGANIZATION_precision": 0.6506024096385542,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8358208955223881,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.8235294117647058,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.31118687987327576,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8448753462603877,
-      "eval_overall_precision": 0.8133333333333334,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 0.9533,
-      "eval_samples_per_second": 196.155,
-      "eval_steps_per_second": 3.147,
+      "eval_loss": 0.29929259419441223,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8387978142076504,
+      "eval_overall_precision": 0.7974025974025974,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3877,
+      "eval_samples_per_second": 482.363,
+      "eval_steps_per_second": 7.738,
       "step": 3392
     },
     {
       "epoch": 33.0,
-      "grad_norm": 0.4719323217868805,
+      "grad_norm": 0.018781477585434914,
       "learning_rate": 3.35e-05,
-      "loss": 0.0023,
+      "loss": 0.003,
       "step": 3498
     },
     {
       "epoch": 33.0,
-      "eval_LOCATION_f1": 0.8476821192052981,
+      "eval_LOCATION_f1": 0.8211920529801324,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8,
-      "eval_LOCATION_recall": 0.9014084507042254,
-      "eval_ORGANIZATION_f1": 0.7703703703703704,
+      "eval_LOCATION_precision": 0.775,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7428571428571429,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7837837837837839,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7073170731707317,
+      "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8620689655172413,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.8333333333333334,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.3178064227104187,
-      "eval_overall_accuracy": 0.9638700290979632,
-      "eval_overall_f1": 0.853185595567867,
-      "eval_overall_precision": 0.8213333333333334,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9706,
-      "eval_samples_per_second": 192.671,
-      "eval_steps_per_second": 3.091,
+      "eval_loss": 0.30042967200279236,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8551724137931035,
+      "eval_overall_precision": 0.8201058201058201,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3707,
+      "eval_samples_per_second": 504.467,
+      "eval_steps_per_second": 8.093,
       "step": 3498
     },
     {
       "epoch": 34.0,
-      "grad_norm": 0.008191280998289585,
+      "grad_norm": 0.05602815002202988,
       "learning_rate": 3.3e-05,
-      "loss": 0.0045,
+      "loss": 0.002,
       "step": 3604
     },
     {
       "epoch": 34.0,
-      "eval_LOCATION_f1": 0.8456375838926176,
+      "eval_LOCATION_f1": 0.8266666666666665,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8076923076923077,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7777777777777778,
+      "eval_LOCATION_precision": 0.7848101265822784,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7746478873239436,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7088607594936709,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.7142857142857143,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9215686274509804,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.9038461538461539,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7837837837837839,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7073170731707317,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2925824224948883,
-      "eval_overall_accuracy": 0.9638700290979632,
-      "eval_overall_f1": 0.8454172366621068,
-      "eval_overall_precision": 0.8046875,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9524,
-      "eval_samples_per_second": 196.348,
-      "eval_steps_per_second": 3.15,
+      "eval_loss": 0.3288598358631134,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.8532235939643347,
+      "eval_overall_precision": 0.8141361256544503,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.378,
+      "eval_samples_per_second": 494.653,
+      "eval_steps_per_second": 7.936,
       "step": 3604
     },
     {
       "epoch": 35.0,
-      "grad_norm": 0.005140832159668207,
+      "grad_norm": 0.006879552733153105,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.0026,
+      "loss": 0.0016,
       "step": 3710
     },
     {
       "epoch": 35.0,
-      "eval_LOCATION_f1": 0.8387096774193548,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7738095238095238,
-      "eval_LOCATION_recall": 0.9154929577464789,
-      "eval_ORGANIZATION_f1": 0.7375,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7883211678832116,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6210526315789474,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.75,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9013157894736842,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7945205479452054,
+      "eval_PERSON_precision": 0.8896103896103896,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.725,
+      "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.78125,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.6944444444444444,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.3151751160621643,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8375165125495376,
-      "eval_overall_precision": 0.7731707317073171,
-      "eval_overall_recall": 0.9135446685878963,
-      "eval_runtime": 0.9517,
-      "eval_samples_per_second": 196.484,
-      "eval_steps_per_second": 3.152,
+      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.3239135146141052,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.853185595567867,
+      "eval_overall_precision": 0.8213333333333334,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3675,
+      "eval_samples_per_second": 508.878,
+      "eval_steps_per_second": 8.164,
       "step": 3710
     },
     {
       "epoch": 36.0,
-      "grad_norm": 0.002234996994957328,
+      "grad_norm": 0.004560971166938543,
       "learning_rate": 3.2000000000000005e-05,
-      "loss": 0.0017,
+      "loss": 0.0024,
       "step": 3816
     },
     {
       "epoch": 36.0,
-      "eval_LOCATION_f1": 0.8356164383561644,
+      "eval_LOCATION_f1": 0.8456375838926176,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8133333333333334,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.8059701492537313,
+      "eval_LOCATION_precision": 0.8076923076923077,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.802919708029197,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.782608695652174,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.7638888888888888,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
+      "eval_PERSON_precision": 0.8853503184713376,
       "eval_PERSON_recall": 0.9266666666666666,
       "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8727272727272727,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8888888888888888,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.27503567934036255,
-      "eval_overall_accuracy": 0.9670223084384093,
-      "eval_overall_f1": 0.8635724331926864,
-      "eval_overall_precision": 0.8434065934065934,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 1.0319,
-      "eval_samples_per_second": 181.221,
-      "eval_steps_per_second": 2.907,
+      "eval_loss": 0.3065287172794342,
+      "eval_overall_accuracy": 0.9662948593598448,
+      "eval_overall_f1": 0.8635097493036211,
+      "eval_overall_precision": 0.8355795148247979,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3687,
+      "eval_samples_per_second": 507.173,
+      "eval_steps_per_second": 8.136,
       "step": 3816
     },
     {
       "epoch": 37.0,
-      "grad_norm": 2.4394583702087402,
+      "grad_norm": 0.0018331906758248806,
       "learning_rate": 3.15e-05,
-      "loss": 0.0016,
+      "loss": 0.0027,
       "step": 3922
     },
     {
       "epoch": 37.0,
-      "eval_LOCATION_f1": 0.8456375838926176,
+      "eval_LOCATION_f1": 0.832214765100671,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8076923076923077,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7755102040816327,
+      "eval_LOCATION_precision": 0.7948717948717948,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.763888888888889,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6951219512195121,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9049180327868852,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_PERSON_precision": 0.8903225806451613,
+      "eval_PERSON_recall": 0.92,
       "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8333333333333334,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.78125,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.3136846423149109,
-      "eval_overall_accuracy": 0.9595053346265762,
-      "eval_overall_f1": 0.8439620081411126,
-      "eval_overall_precision": 0.7974358974358975,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9561,
-      "eval_samples_per_second": 195.586,
-      "eval_steps_per_second": 3.138,
+      "eval_TIME_precision": 0.8,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.30563417077064514,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8461538461538461,
+      "eval_overall_precision": 0.8083989501312336,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3699,
+      "eval_samples_per_second": 505.544,
+      "eval_steps_per_second": 8.11,
       "step": 3922
     },
     {
       "epoch": 38.0,
-      "grad_norm": 0.02068762294948101,
+      "grad_norm": 0.002592272823676467,
       "learning_rate": 3.1e-05,
-      "loss": 0.002,
+      "loss": 0.0015,
       "step": 4028
     },
     {
       "epoch": 38.0,
-      "eval_LOCATION_f1": 0.8163265306122449,
+      "eval_LOCATION_f1": 0.7924528301886792,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7894736842105263,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7619047619047619,
+      "eval_LOCATION_precision": 0.7159090909090909,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7714285714285715,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6829268292682927,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.72,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
+      "eval_PERSON_precision": 0.879746835443038,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7837837837837839,
+      "eval_QUANTITY_f1": 0.8333333333333333,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7073170731707317,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8333333333333334,
+      "eval_QUANTITY_precision": 0.7692307692307693,
+      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.78125,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.31167086958885193,
-      "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.8431105047748976,
-      "eval_overall_precision": 0.8005181347150259,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9533,
-      "eval_samples_per_second": 196.17,
-      "eval_steps_per_second": 3.147,
+      "eval_TIME_precision": 0.8,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.33764445781707764,
+      "eval_overall_accuracy": 0.9599903006789525,
+      "eval_overall_f1": 0.841248303934871,
+      "eval_overall_precision": 0.7948717948717948,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3802,
+      "eval_samples_per_second": 491.847,
+      "eval_steps_per_second": 7.891,
       "step": 4028
     },
     {
       "epoch": 39.0,
-      "grad_norm": 0.004659323953092098,
+      "grad_norm": 0.0023491207975894213,
       "learning_rate": 3.05e-05,
-      "loss": 0.0031,
+      "loss": 0.0039,
       "step": 4134
     },
     {
       "epoch": 39.0,
-      "eval_LOCATION_f1": 0.8421052631578947,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7901234567901234,
+      "eval_LOCATION_precision": 0.7710843373493976,
       "eval_LOCATION_recall": 0.9014084507042254,
-      "eval_ORGANIZATION_f1": 0.7682119205298013,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6744186046511628,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
+      "eval_PERSON_precision": 0.879746835443038,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.823529411764706,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8,
+      "eval_QUANTITY_precision": 0.717948717948718,
       "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.819672131147541,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7575757575757576,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.3022630512714386,
-      "eval_overall_accuracy": 0.9609602327837051,
-      "eval_overall_f1": 0.8521031207598372,
-      "eval_overall_precision": 0.8051282051282052,
-      "eval_overall_recall": 0.9048991354466859,
-      "eval_runtime": 0.9507,
-      "eval_samples_per_second": 196.705,
-      "eval_steps_per_second": 3.156,
+      "eval_TIME_precision": 0.8,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.29880788922309875,
+      "eval_overall_accuracy": 0.9633850630455868,
+      "eval_overall_f1": 0.8415300546448087,
+      "eval_overall_precision": 0.8,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3813,
+      "eval_samples_per_second": 490.458,
+      "eval_steps_per_second": 7.868,
       "step": 4134
     },
     {
       "epoch": 40.0,
-      "grad_norm": 0.5910449624061584,
+      "grad_norm": 0.15171128511428833,
       "learning_rate": 3e-05,
-      "loss": 0.0018,
+      "loss": 0.0022,
       "step": 4240
     },
     {
       "epoch": 40.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8205128205128205,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7763157894736842,
+      "eval_LOCATION_precision": 0.7529411764705882,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6781609195402298,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9180327868852459,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
-      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_recall": 0.9333333333333333,
       "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7567567567567568,
       "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": 0.31180375814437866,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8416779431664412,
-      "eval_overall_precision": 0.7933673469387755,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9409,
-      "eval_samples_per_second": 198.747,
-      "eval_steps_per_second": 3.188,
+      "eval_loss": 0.30401691794395447,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8512396694214875,
+      "eval_overall_precision": 0.8153034300791556,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3795,
+      "eval_samples_per_second": 492.762,
+      "eval_steps_per_second": 7.905,
       "step": 4240
     },
     {
       "epoch": 41.0,
-      "grad_norm": 0.0014322897186502814,
+      "grad_norm": 0.0015246509574353695,
       "learning_rate": 2.95e-05,
-      "loss": 0.0022,
+      "loss": 0.0014,
       "step": 4346
     },
     {
       "epoch": 41.0,
-      "eval_LOCATION_f1": 0.8187919463087248,
+      "eval_LOCATION_f1": 0.8413793103448276,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.782051282051282,
+      "eval_LOCATION_precision": 0.8243243243243243,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7832167832167831,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.717948717948718,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9036544850498339,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9180327868852459,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9006622516556292,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.2662249207496643,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.8515950069348127,
-      "eval_overall_precision": 0.820855614973262,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9578,
-      "eval_samples_per_second": 195.232,
-      "eval_steps_per_second": 3.132,
+      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.3144817352294922,
+      "eval_overall_accuracy": 0.9655674102812803,
+      "eval_overall_f1": 0.8603351955307261,
+      "eval_overall_precision": 0.8346883468834688,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3689,
+      "eval_samples_per_second": 506.863,
+      "eval_steps_per_second": 8.131,
       "step": 4346
     },
     {
       "epoch": 42.0,
-      "grad_norm": 0.0011240842286497355,
+      "grad_norm": 0.0009719623485580087,
       "learning_rate": 2.9e-05,
-      "loss": 0.0006,
+      "loss": 0.0032,
       "step": 4452
     },
     {
       "epoch": 42.0,
-      "eval_LOCATION_f1": 0.7866666666666666,
+      "eval_LOCATION_f1": 0.8152866242038216,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7468354430379747,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7712418300653594,
+      "eval_LOCATION_precision": 0.7441860465116279,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7448275862068966,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6704545454545454,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.675,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9067524115755627,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
+      "eval_PERSON_precision": 0.8757763975155279,
+      "eval_PERSON_recall": 0.94,
       "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_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3214167058467865,
+      "eval_loss": 0.3701828420162201,
       "eval_overall_accuracy": 0.9597478176527643,
-      "eval_overall_f1": 0.8373983739837398,
-      "eval_overall_precision": 0.7902813299232737,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9589,
-      "eval_samples_per_second": 195.011,
-      "eval_steps_per_second": 3.129,
+      "eval_overall_f1": 0.8375838926174496,
+      "eval_overall_precision": 0.7839195979899497,
+      "eval_overall_recall": 0.899135446685879,
+      "eval_runtime": 0.366,
+      "eval_samples_per_second": 510.969,
+      "eval_steps_per_second": 8.197,
       "step": 4452
     },
     {
       "epoch": 43.0,
-      "grad_norm": 0.0019562048837542534,
+      "grad_norm": 1.7098315954208374,
       "learning_rate": 2.8499999999999998e-05,
-      "loss": 0.0011,
+      "loss": 0.0022,
       "step": 4558
     },
     {
       "epoch": 43.0,
-      "eval_LOCATION_f1": 0.7712418300653594,
+      "eval_LOCATION_f1": 0.8076923076923077,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7195121951219512,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7785234899328859,
+      "eval_LOCATION_precision": 0.7411764705882353,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7857142857142856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6904761904761905,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
+      "eval_PERSON_precision": 0.8853503184713376,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8169014084507042,
+      "eval_QUANTITY_f1": 0.7777777777777778,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7631578947368421,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3223561942577362,
-      "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.8408163265306122,
-      "eval_overall_precision": 0.7963917525773195,
+      "eval_loss": 0.3241700828075409,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.8419618528610354,
+      "eval_overall_precision": 0.7984496124031008,
       "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9419,
-      "eval_samples_per_second": 198.529,
-      "eval_steps_per_second": 3.185,
+      "eval_runtime": 0.3704,
+      "eval_samples_per_second": 504.841,
+      "eval_steps_per_second": 8.099,
       "step": 4558
     },
     {
       "epoch": 44.0,
-      "grad_norm": 0.004401817452162504,
+      "grad_norm": 0.01855984516441822,
       "learning_rate": 2.8000000000000003e-05,
-      "loss": 0.0015,
+      "loss": 0.0012,
       "step": 4664
     },
     {
       "epoch": 44.0,
-      "eval_LOCATION_f1": 0.8344370860927152,
+      "eval_LOCATION_f1": 0.8076923076923077,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_precision": 0.7411764705882353,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7402597402597403,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6404494382022472,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9013157894736842,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8896103896103896,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "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": 0.30491575598716736,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8428184281842819,
-      "eval_overall_precision": 0.7953964194373402,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9628,
-      "eval_samples_per_second": 194.222,
-      "eval_steps_per_second": 3.116,
+      "eval_loss": 0.3312273621559143,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8419618528610354,
+      "eval_overall_precision": 0.7984496124031008,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3782,
+      "eval_samples_per_second": 494.494,
+      "eval_steps_per_second": 7.933,
       "step": 4664
     },
     {
       "epoch": 45.0,
-      "grad_norm": 0.032622966915369034,
+      "grad_norm": 0.00311101833358407,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.0023,
+      "loss": 0.0026,
       "step": 4770
     },
     {
       "epoch": 45.0,
-      "eval_LOCATION_f1": 0.8187919463087248,
+      "eval_LOCATION_f1": 0.7870967741935485,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.782051282051282,
+      "eval_LOCATION_precision": 0.7261904761904762,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.763888888888889,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6962025316455697,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8933333333333333,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8933333333333333,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.8529411764705883,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8421052631578947,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8275862068965517,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.26732492446899414,
-      "eval_overall_accuracy": 0.9648399612027158,
-      "eval_overall_f1": 0.8440111420612814,
-      "eval_overall_precision": 0.816711590296496,
-      "eval_overall_recall": 0.8731988472622478,
-      "eval_runtime": 1.0306,
-      "eval_samples_per_second": 181.45,
-      "eval_steps_per_second": 2.911,
+      "eval_loss": 0.3425029516220093,
+      "eval_overall_accuracy": 0.9612027158098934,
+      "eval_overall_f1": 0.8415300546448087,
+      "eval_overall_precision": 0.8,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3681,
+      "eval_samples_per_second": 508.054,
+      "eval_steps_per_second": 8.151,
       "step": 4770
     },
     {
       "epoch": 46.0,
-      "grad_norm": 0.009263483807444572,
+      "grad_norm": 0.004660445265471935,
       "learning_rate": 2.7000000000000002e-05,
-      "loss": 0.0015,
+      "loss": 0.0032,
       "step": 4876
     },
     {
       "epoch": 46.0,
-      "eval_LOCATION_f1": 0.8435374149659864,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8157894736842105,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7769784172661871,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7462686567164178,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7297297297297297,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.913907284768212,
+      "eval_ORGANIZATION_precision": 0.7246376811594203,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9078947368421053,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "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": 0.2788328230381012,
-      "eval_overall_accuracy": 0.9662948593598448,
-      "eval_overall_f1": 0.8587412587412588,
-      "eval_overall_precision": 0.8342391304347826,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9441,
-      "eval_samples_per_second": 198.082,
-      "eval_steps_per_second": 3.178,
+      "eval_loss": 0.30848801136016846,
+      "eval_overall_accuracy": 0.962172647914646,
+      "eval_overall_f1": 0.8441379310344828,
+      "eval_overall_precision": 0.8095238095238095,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3698,
+      "eval_samples_per_second": 505.735,
+      "eval_steps_per_second": 8.113,
       "step": 4876
     },
     {
       "epoch": 47.0,
-      "grad_norm": 0.016023587435483932,
+      "grad_norm": 0.2071634829044342,
       "learning_rate": 2.6500000000000004e-05,
-      "loss": 0.0019,
+      "loss": 0.0013,
       "step": 4982
     },
     {
       "epoch": 47.0,
-      "eval_LOCATION_f1": 0.8513513513513514,
+      "eval_LOCATION_f1": 0.8051948051948051,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8181818181818182,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.8235294117647058,
+      "eval_LOCATION_precision": 0.7469879518072289,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7517730496453899,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7887323943661971,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9169435215946843,
+      "eval_ORGANIZATION_precision": 0.6973684210526315,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9139072847682119,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "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_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.717948717948718,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.2903594672679901,
-      "eval_overall_accuracy": 0.9670223084384093,
-      "eval_overall_f1": 0.8671328671328672,
-      "eval_overall_precision": 0.842391304347826,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9543,
-      "eval_samples_per_second": 195.955,
-      "eval_steps_per_second": 3.144,
+      "eval_loss": 0.3372473120689392,
+      "eval_overall_accuracy": 0.9602327837051406,
+      "eval_overall_f1": 0.8353741496598639,
+      "eval_overall_precision": 0.7912371134020618,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3668,
+      "eval_samples_per_second": 509.807,
+      "eval_steps_per_second": 8.179,
       "step": 4982
     },
     {
       "epoch": 48.0,
-      "grad_norm": 0.0013861858751624823,
+      "grad_norm": 1.3425573110580444,
       "learning_rate": 2.6000000000000002e-05,
-      "loss": 0.0016,
+      "loss": 0.0015,
       "step": 5088
     },
     {
       "epoch": 48.0,
-      "eval_LOCATION_f1": 0.8,
+      "eval_LOCATION_f1": 0.7672955974842768,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.759493670886076,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7755102040816327,
+      "eval_LOCATION_precision": 0.6931818181818182,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.7361111111111112,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6951219512195121,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9174917491749175,
+      "eval_ORGANIZATION_precision": 0.6708860759493671,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9084967320261438,
-      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_recall": 0.92,
       "eval_QUANTITY_f1": 0.8055555555555556,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7435897435897436,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.30562856793403625,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8454172366621068,
-      "eval_overall_precision": 0.8046875,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.953,
-      "eval_samples_per_second": 196.222,
-      "eval_steps_per_second": 3.148,
+      "eval_loss": 0.3451178967952728,
+      "eval_overall_accuracy": 0.9592628516003879,
+      "eval_overall_f1": 0.8209959623149394,
+      "eval_overall_precision": 0.7702020202020202,
+      "eval_overall_recall": 0.8789625360230547,
+      "eval_runtime": 0.3675,
+      "eval_samples_per_second": 508.868,
+      "eval_steps_per_second": 8.164,
       "step": 5088
     },
     {
       "epoch": 49.0,
-      "grad_norm": 0.004494914319366217,
+      "grad_norm": 0.0004966954002156854,
       "learning_rate": 2.5500000000000003e-05,
-      "loss": 0.0021,
+      "loss": 0.001,
       "step": 5194
     },
     {
       "epoch": 49.0,
-      "eval_LOCATION_f1": 0.8266666666666665,
+      "eval_LOCATION_f1": 0.7948717948717949,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7848101265822784,
+      "eval_LOCATION_precision": 0.7294117647058823,
       "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.913907284768212,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9078947368421053,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8474576271186439,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8064516129032258,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.30186086893081665,
-      "eval_overall_accuracy": 0.9643549951503395,
-      "eval_overall_f1": 0.8583333333333334,
-      "eval_overall_precision": 0.8284182305630027,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9632,
-      "eval_samples_per_second": 194.143,
-      "eval_steps_per_second": 3.115,
+      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.3490431308746338,
+      "eval_overall_accuracy": 0.9614451988360815,
+      "eval_overall_f1": 0.834924965893588,
+      "eval_overall_precision": 0.7927461139896373,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.368,
+      "eval_samples_per_second": 508.154,
+      "eval_steps_per_second": 8.152,
       "step": 5194
     },
     {
       "epoch": 50.0,
-      "grad_norm": 0.07165616750717163,
+      "grad_norm": 0.002857360290363431,
       "learning_rate": 2.5e-05,
-      "loss": 0.0013,
+      "loss": 0.0019,
       "step": 5300
     },
     {
       "epoch": 50.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.8079470198675496,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.75,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.76056338028169,
+      "eval_LOCATION_precision": 0.7625,
+      "eval_LOCATION_recall": 0.8591549295774648,
+      "eval_ORGANIZATION_f1": 0.725925925925926,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7012987012987013,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.7,
+      "eval_ORGANIZATION_recall": 0.7538461538461538,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8115942028985507,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7777777777777778,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3192538917064667,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8395061728395062,
-      "eval_overall_precision": 0.8010471204188482,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9477,
-      "eval_samples_per_second": 197.324,
-      "eval_steps_per_second": 3.166,
+      "eval_loss": 0.3414745330810547,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.838888888888889,
+      "eval_overall_precision": 0.8096514745308311,
+      "eval_overall_recall": 0.8703170028818443,
+      "eval_runtime": 0.3687,
+      "eval_samples_per_second": 507.166,
+      "eval_steps_per_second": 8.136,
       "step": 5300
     },
     {
       "epoch": 51.0,
-      "grad_norm": 0.0017693530535325408,
+      "grad_norm": 0.005967956036329269,
       "learning_rate": 2.45e-05,
-      "loss": 0.0013,
+      "loss": 0.0007,
       "step": 5406
     },
     {
       "epoch": 51.0,
-      "eval_LOCATION_f1": 0.8187919463087248,
+      "eval_LOCATION_f1": 0.8104575163398693,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.782051282051282,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7808219178082192,
+      "eval_LOCATION_precision": 0.7560975609756098,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7647058823529411,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7037037037037037,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_ORGANIZATION_precision": 0.7323943661971831,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8405797101449276,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_precision": 0.8055555555555556,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "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": 0.3206080496311188,
+      "eval_loss": 0.3594156801700592,
       "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.850480109739369,
-      "eval_overall_precision": 0.8115183246073299,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9406,
-      "eval_samples_per_second": 198.804,
-      "eval_steps_per_second": 3.189,
+      "eval_overall_f1": 0.8445667125171938,
+      "eval_overall_precision": 0.8078947368421052,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.368,
+      "eval_samples_per_second": 508.084,
+      "eval_steps_per_second": 8.151,
       "step": 5406
     },
     {
       "epoch": 52.0,
-      "grad_norm": 0.077069491147995,
+      "grad_norm": 0.000697598559781909,
       "learning_rate": 2.4e-05,
-      "loss": 0.0014,
+      "loss": 0.0004,
       "step": 5512
     },
     {
       "epoch": 52.0,
-      "eval_LOCATION_f1": 0.8299319727891157,
+      "eval_LOCATION_f1": 0.8025477707006369,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8026315789473685,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7801418439716312,
+      "eval_LOCATION_precision": 0.7325581395348837,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7769784172661871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7236842105263158,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.7297297297297297,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7945205479452054,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_precision": 0.725,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "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": 0.29891258478164673,
-      "eval_overall_accuracy": 0.9653249272550921,
-      "eval_overall_f1": 0.8567454798331016,
-      "eval_overall_precision": 0.8279569892473119,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9575,
-      "eval_samples_per_second": 195.296,
-      "eval_steps_per_second": 3.133,
+      "eval_loss": 0.3847062885761261,
+      "eval_overall_accuracy": 0.9604752667313288,
+      "eval_overall_f1": 0.841248303934871,
+      "eval_overall_precision": 0.7948717948717948,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3671,
+      "eval_samples_per_second": 509.355,
+      "eval_steps_per_second": 8.171,
       "step": 5512
     },
     {
       "epoch": 53.0,
-      "grad_norm": 0.002096625277772546,
+      "grad_norm": 0.001760791172273457,
       "learning_rate": 2.35e-05,
-      "loss": 0.0007,
+      "loss": 0.0012,
       "step": 5618
     },
     {
       "epoch": 53.0,
-      "eval_LOCATION_f1": 0.8194444444444445,
+      "eval_LOCATION_f1": 0.8333333333333333,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8082191780821918,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.821917808219178,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.7199999999999999,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.75,
+      "eval_ORGANIZATION_recall": 0.6923076923076923,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
+      "eval_PERSON_precision": 0.8726114649681529,
       "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8405797101449276,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8055555555555556,
+      "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8928571428571429,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8928571428571429,
-      "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.30264776945114136,
-      "eval_overall_accuracy": 0.9672647914645974,
-      "eval_overall_f1": 0.8579465541490857,
-      "eval_overall_precision": 0.8379120879120879,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 0.951,
-      "eval_samples_per_second": 196.632,
-      "eval_steps_per_second": 3.155,
+      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_recall": 0.8571428571428571,
+      "eval_loss": 0.33592331409454346,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8404558404558404,
+      "eval_overall_precision": 0.8309859154929577,
+      "eval_overall_recall": 0.8501440922190202,
+      "eval_runtime": 0.3683,
+      "eval_samples_per_second": 507.672,
+      "eval_steps_per_second": 8.144,
       "step": 5618
     },
     {
       "epoch": 54.0,
-      "grad_norm": 0.001791234826669097,
+      "grad_norm": 0.017065823078155518,
       "learning_rate": 2.3000000000000003e-05,
-      "loss": 0.0014,
+      "loss": 0.0018,
       "step": 5724
     },
     {
       "epoch": 54.0,
-      "eval_LOCATION_f1": 0.8243243243243243,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7922077922077922,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.8055555555555556,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7851851851851851,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7341772151898734,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.8983606557377048,
+      "eval_ORGANIZATION_precision": 0.7571428571428571,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9019607843137256,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8838709677419355,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8529411764705883,
+      "eval_PERSON_precision": 0.8846153846153846,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8771929824561403,
+      "eval_TIME_f1": 0.8620689655172413,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8620689655172413,
+      "eval_TIME_precision": 0.8333333333333334,
       "eval_TIME_recall": 0.8928571428571429,
-      "eval_loss": 0.305812805891037,
-      "eval_overall_accuracy": 0.9670223084384093,
-      "eval_overall_f1": 0.8587257617728532,
-      "eval_overall_precision": 0.8266666666666667,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9512,
-      "eval_samples_per_second": 196.589,
-      "eval_steps_per_second": 3.154,
+      "eval_loss": 0.3532910943031311,
+      "eval_overall_accuracy": 0.9633850630455868,
+      "eval_overall_f1": 0.853185595567867,
+      "eval_overall_precision": 0.8213333333333334,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3682,
+      "eval_samples_per_second": 507.911,
+      "eval_steps_per_second": 8.148,
       "step": 5724
     },
     {
       "epoch": 55.0,
-      "grad_norm": 0.0006963410996831954,
+      "grad_norm": 0.0010331659577786922,
       "learning_rate": 2.25e-05,
-      "loss": 0.0013,
+      "loss": 0.0006,
       "step": 5830
     },
     {
       "epoch": 55.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7826086956521738,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7883211678832116,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7397260273972602,
+      "eval_ORGANIZATION_precision": 0.75,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7714285714285715,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7297297297297297,
-      "eval_QUANTITY_recall": 0.8181818181818182,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.31985101103782654,
-      "eval_overall_accuracy": 0.9619301648884578,
-      "eval_overall_f1": 0.8448753462603877,
-      "eval_overall_precision": 0.8133333333333334,
-      "eval_overall_recall": 0.8789625360230547,
-      "eval_runtime": 0.9536,
-      "eval_samples_per_second": 196.106,
-      "eval_steps_per_second": 3.146,
+      "eval_loss": 0.35429704189300537,
+      "eval_overall_accuracy": 0.9626576139670223,
+      "eval_overall_f1": 0.8469945355191256,
+      "eval_overall_precision": 0.8051948051948052,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3678,
+      "eval_samples_per_second": 508.392,
+      "eval_steps_per_second": 8.156,
       "step": 5830
     },
     {
       "epoch": 56.0,
-      "grad_norm": 0.004016083665192127,
+      "grad_norm": 0.0061631957069039345,
       "learning_rate": 2.2000000000000003e-05,
-      "loss": 0.002,
+      "loss": 0.0012,
       "step": 5936
     },
     {
       "epoch": 56.0,
-      "eval_LOCATION_f1": 0.8266666666666665,
+      "eval_LOCATION_f1": 0.8289473684210527,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7971014492753623,
+      "eval_LOCATION_precision": 0.7777777777777778,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7938931297709925,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7534246575342466,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_ORGANIZATION_precision": 0.7878787878787878,
+      "eval_ORGANIZATION_recall": 0.8,
       "eval_PERSON_f1": 0.9084967320261438,
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.8910256410256411,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8055555555555556,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7435897435897436,
+      "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.33322784304618835,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8524137931034482,
-      "eval_overall_precision": 0.8174603174603174,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9475,
-      "eval_samples_per_second": 197.363,
-      "eval_steps_per_second": 3.166,
+      "eval_loss": 0.3279174566268921,
+      "eval_overall_accuracy": 0.9653249272550921,
+      "eval_overall_f1": 0.8539638386648122,
+      "eval_overall_precision": 0.8252688172043011,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3765,
+      "eval_samples_per_second": 496.69,
+      "eval_steps_per_second": 7.968,
       "step": 5936
     },
     {
       "epoch": 57.0,
-      "grad_norm": 0.002730452921241522,
+      "grad_norm": 1.2823433876037598,
       "learning_rate": 2.15e-05,
       "loss": 0.0006,
       "step": 6042
     },
     {
       "epoch": 57.0,
-      "eval_LOCATION_f1": 0.8211920529801324,
+      "eval_LOCATION_f1": 0.832214765100671,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.775,
+      "eval_LOCATION_precision": 0.7948717948717948,
       "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7862068965517242,
+      "eval_ORGANIZATION_f1": 0.8091603053435115,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7125,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.803030303030303,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9019607843137256,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_PERSON_precision": 0.8846153846153846,
+      "eval_PERSON_recall": 0.92,
       "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_TIME_f1": 0.819672131147541,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3458655774593353,
-      "eval_overall_accuracy": 0.9612027158098934,
-      "eval_overall_f1": 0.8497267759562841,
-      "eval_overall_precision": 0.8077922077922078,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9455,
-      "eval_samples_per_second": 197.779,
-      "eval_steps_per_second": 3.173,
+      "eval_TIME_precision": 0.7575757575757576,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.3345893621444702,
+      "eval_overall_accuracy": 0.9658098933074685,
+      "eval_overall_f1": 0.8551532033426185,
+      "eval_overall_precision": 0.8274932614555256,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3676,
+      "eval_samples_per_second": 508.689,
+      "eval_steps_per_second": 8.161,
       "step": 6042
     },
     {
       "epoch": 58.0,
-      "grad_norm": 0.0037386000622063875,
+      "grad_norm": 0.0004909472190774977,
       "learning_rate": 2.1e-05,
-      "loss": 0.0014,
+      "loss": 0.0012,
       "step": 6148
     },
     {
       "epoch": 58.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8187919463087248,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
+      "eval_LOCATION_precision": 0.782051282051282,
       "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.802721088435374,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7195121951219512,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9055374592833875,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
+      "eval_PERSON_precision": 0.8853503184713376,
+      "eval_PERSON_recall": 0.9266666666666666,
       "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7631578947368421,
@@ -2260,1662 +2260,1662 @@
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.32678091526031494,
-      "eval_overall_accuracy": 0.9614451988360815,
-      "eval_overall_f1": 0.8485675306957708,
-      "eval_overall_precision": 0.805699481865285,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9604,
-      "eval_samples_per_second": 194.708,
-      "eval_steps_per_second": 3.124,
+      "eval_loss": 0.3291900157928467,
+      "eval_overall_accuracy": 0.9636275460717749,
+      "eval_overall_f1": 0.8496551724137931,
+      "eval_overall_precision": 0.8148148148148148,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3782,
+      "eval_samples_per_second": 494.476,
+      "eval_steps_per_second": 7.933,
       "step": 6148
     },
     {
       "epoch": 59.0,
-      "grad_norm": 0.0023347027599811554,
+      "grad_norm": 0.0017112432979047298,
       "learning_rate": 2.05e-05,
-      "loss": 0.0012,
+      "loss": 0.0011,
       "step": 6254
     },
     {
       "epoch": 59.0,
-      "eval_LOCATION_f1": 0.8181818181818181,
+      "eval_LOCATION_f1": 0.8051948051948051,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7590361445783133,
-      "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7631578947368421,
+      "eval_LOCATION_precision": 0.7469879518072289,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.8,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9042904290429042,
+      "eval_ORGANIZATION_precision": 0.7714285714285715,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8990228013029316,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8954248366013072,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.8789808917197452,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3488438129425049,
-      "eval_overall_accuracy": 0.9604752667313288,
-      "eval_overall_f1": 0.8455284552845529,
-      "eval_overall_precision": 0.7979539641943734,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9526,
-      "eval_samples_per_second": 196.312,
-      "eval_steps_per_second": 3.149,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.34627169370651245,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8473177441540578,
+      "eval_overall_precision": 0.8105263157894737,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3665,
+      "eval_samples_per_second": 510.28,
+      "eval_steps_per_second": 8.186,
       "step": 6254
     },
     {
       "epoch": 60.0,
-      "grad_norm": 0.0007857033051550388,
+      "grad_norm": 0.001461641164496541,
       "learning_rate": 2e-05,
-      "loss": 0.0015,
+      "loss": 0.0006,
       "step": 6360
     },
     {
       "epoch": 60.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.8051948051948051,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7724137931034483,
+      "eval_LOCATION_precision": 0.7469879518072289,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7846153846153847,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9108910891089109,
+      "eval_ORGANIZATION_precision": 0.7846153846153846,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9084967320261438,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9019607843137255,
-      "eval_PERSON_recall": 0.92,
+      "eval_PERSON_precision": 0.8910256410256411,
+      "eval_PERSON_recall": 0.9266666666666666,
       "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7837837837837838,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8813559322033899,
+      "eval_TIME_f1": 0.8474576271186439,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8387096774193549,
-      "eval_TIME_recall": 0.9285714285714286,
-      "eval_loss": 0.3345150947570801,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8528198074277854,
-      "eval_overall_precision": 0.8157894736842105,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9473,
-      "eval_samples_per_second": 197.404,
-      "eval_steps_per_second": 3.167,
+      "eval_TIME_precision": 0.8064516129032258,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.3555537760257721,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.851182197496523,
+      "eval_overall_precision": 0.8225806451612904,
+      "eval_overall_recall": 0.8818443804034583,
+      "eval_runtime": 0.3827,
+      "eval_samples_per_second": 488.657,
+      "eval_steps_per_second": 7.839,
       "step": 6360
     },
     {
       "epoch": 61.0,
-      "grad_norm": 0.0016401028260588646,
+      "grad_norm": 0.00039941485738381743,
       "learning_rate": 1.9500000000000003e-05,
-      "loss": 0.0005,
+      "loss": 0.0008,
       "step": 6466
     },
     {
       "epoch": 61.0,
-      "eval_LOCATION_f1": 0.8157894736842104,
+      "eval_LOCATION_f1": 0.7843137254901961,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7654320987654321,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.8,
+      "eval_LOCATION_precision": 0.7317073170731707,
+      "eval_LOCATION_recall": 0.8450704225352113,
+      "eval_ORGANIZATION_f1": 0.8059701492537313,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7466666666666667,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_ORGANIZATION_precision": 0.782608695652174,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9185667752442996,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.8980891719745223,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3476317226886749,
-      "eval_overall_accuracy": 0.9624151309408342,
-      "eval_overall_f1": 0.8539944903581268,
-      "eval_overall_precision": 0.8179419525065963,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9315,
-      "eval_samples_per_second": 200.75,
-      "eval_steps_per_second": 3.221,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.3484286963939667,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8496551724137931,
+      "eval_overall_precision": 0.8148148148148148,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3665,
+      "eval_samples_per_second": 510.238,
+      "eval_steps_per_second": 8.186,
       "step": 6466
     },
     {
       "epoch": 62.0,
-      "grad_norm": 0.02758892811834812,
+      "grad_norm": 0.06066470965743065,
       "learning_rate": 1.9e-05,
-      "loss": 0.001,
+      "loss": 0.0013,
       "step": 6572
     },
     {
       "epoch": 62.0,
-      "eval_LOCATION_f1": 0.8211920529801324,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.775,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.8028169014084509,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.8059701492537313,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7402597402597403,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_ORGANIZATION_precision": 0.782608695652174,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7936507936507937,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3546055257320404,
-      "eval_overall_accuracy": 0.9633850630455868,
-      "eval_overall_f1": 0.8539944903581268,
-      "eval_overall_precision": 0.8179419525065963,
+      "eval_TIME_precision": 0.7142857142857143,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.35640770196914673,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
       "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9503,
-      "eval_samples_per_second": 196.775,
-      "eval_steps_per_second": 3.157,
+      "eval_runtime": 0.3713,
+      "eval_samples_per_second": 503.58,
+      "eval_steps_per_second": 8.079,
       "step": 6572
     },
     {
       "epoch": 63.0,
-      "grad_norm": 0.32324543595314026,
+      "grad_norm": 0.3191024959087372,
       "learning_rate": 1.85e-05,
-      "loss": 0.0008,
+      "loss": 0.0015,
       "step": 6678
     },
     {
       "epoch": 63.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.8152866242038216,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7862068965517242,
+      "eval_LOCATION_precision": 0.7441860465116279,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7125,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.912621359223301,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8867924528301887,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.819672131147541,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3560656011104584,
-      "eval_overall_accuracy": 0.9616876818622696,
-      "eval_overall_f1": 0.8481532147742818,
-      "eval_overall_precision": 0.8072916666666666,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9525,
-      "eval_samples_per_second": 196.325,
-      "eval_steps_per_second": 3.15,
+      "eval_TIME_precision": 0.7575757575757576,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.34782156348228455,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8451086956521738,
+      "eval_overall_precision": 0.7994858611825193,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3684,
+      "eval_samples_per_second": 507.601,
+      "eval_steps_per_second": 8.143,
       "step": 6678
     },
     {
       "epoch": 64.0,
-      "grad_norm": 0.0021076402626931667,
+      "grad_norm": 0.0011131856590509415,
       "learning_rate": 1.8e-05,
-      "loss": 0.0012,
+      "loss": 0.0007,
       "step": 6784
     },
     {
       "epoch": 64.0,
-      "eval_LOCATION_f1": 0.8344370860927152,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_precision": 0.7682926829268293,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7887323943661972,
+      "eval_ORGANIZATION_f1": 0.7819548872180452,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7272727272727273,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.7647058823529411,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9180327868852459,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.33117640018463135,
-      "eval_overall_accuracy": 0.9624151309408342,
-      "eval_overall_f1": 0.8567493112947657,
-      "eval_overall_precision": 0.820580474934037,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9576,
-      "eval_samples_per_second": 195.279,
-      "eval_steps_per_second": 3.133,
-      "step": 6784
-    },
-    {
-      "epoch": 65.0,
-      "grad_norm": 0.27350062131881714,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.3234313130378723,
+      "eval_overall_accuracy": 0.9643549951503395,
+      "eval_overall_f1": 0.8543689320388349,
+      "eval_overall_precision": 0.8235294117647058,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3804,
+      "eval_samples_per_second": 491.525,
+      "eval_steps_per_second": 7.885,
+      "step": 6784
+    },
+    {
+      "epoch": 65.0,
+      "grad_norm": 0.29352253675460815,
       "learning_rate": 1.75e-05,
-      "loss": 0.0017,
+      "loss": 0.0007,
       "step": 6890
     },
     {
       "epoch": 65.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8104575163398693,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.7560975609756098,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.819672131147541,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3442927598953247,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8461538461538461,
-      "eval_overall_precision": 0.8083989501312336,
+      "eval_TIME_precision": 0.7575757575757576,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.342242032289505,
+      "eval_overall_accuracy": 0.9624151309408342,
+      "eval_overall_f1": 0.8449931412894376,
+      "eval_overall_precision": 0.806282722513089,
       "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.953,
-      "eval_samples_per_second": 196.22,
-      "eval_steps_per_second": 3.148,
+      "eval_runtime": 0.3681,
+      "eval_samples_per_second": 508.007,
+      "eval_steps_per_second": 8.15,
       "step": 6890
     },
     {
       "epoch": 66.0,
-      "grad_norm": 0.7737051248550415,
+      "grad_norm": 0.0007412360864691436,
       "learning_rate": 1.7000000000000003e-05,
-      "loss": 0.0005,
+      "loss": 0.0004,
       "step": 6996
     },
     {
       "epoch": 66.0,
-      "eval_LOCATION_f1": 0.7763157894736843,
+      "eval_LOCATION_f1": 0.8421052631578947,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7283950617283951,
-      "eval_LOCATION_recall": 0.8309859154929577,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.7901234567901234,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7727272727272727,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9055374592833875,
+      "eval_ORGANIZATION_precision": 0.7611940298507462,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8853503184713376,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8169014084507042,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7631578947368421,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7868852459016394,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8333333333333334,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7272727272727273,
-      "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3465050756931305,
-      "eval_overall_accuracy": 0.9602327837051406,
-      "eval_overall_f1": 0.8383561643835616,
-      "eval_overall_precision": 0.7989556135770235,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9746,
-      "eval_samples_per_second": 191.87,
-      "eval_steps_per_second": 3.078,
+      "eval_TIME_precision": 0.78125,
+      "eval_TIME_recall": 0.8928571428571429,
+      "eval_loss": 0.3538872301578522,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8543689320388349,
+      "eval_overall_precision": 0.8235294117647058,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.374,
+      "eval_samples_per_second": 499.972,
+      "eval_steps_per_second": 8.021,
       "step": 6996
     },
     {
       "epoch": 67.0,
-      "grad_norm": 0.0003085023781750351,
+      "grad_norm": 0.00021199545881245285,
       "learning_rate": 1.65e-05,
-      "loss": 0.0004,
+      "loss": 0.0005,
       "step": 7102
     },
     {
       "epoch": 67.0,
-      "eval_LOCATION_f1": 0.8266666666666665,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.8115942028985507,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7761194029850746,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7671232876712328,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9114754098360656,
+      "eval_ORGANIZATION_precision": 0.7536231884057971,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.896774193548387,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3288689851760864,
-      "eval_overall_accuracy": 0.9638700290979632,
-      "eval_overall_f1": 0.8591160220994475,
-      "eval_overall_precision": 0.8249336870026526,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9508,
-      "eval_samples_per_second": 196.673,
-      "eval_steps_per_second": 3.155,
+      "eval_loss": 0.3606233596801758,
+      "eval_overall_accuracy": 0.9648399612027158,
+      "eval_overall_f1": 0.8551724137931035,
+      "eval_overall_precision": 0.8201058201058201,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3675,
+      "eval_samples_per_second": 508.788,
+      "eval_steps_per_second": 8.162,
       "step": 7102
     },
     {
       "epoch": 68.0,
-      "grad_norm": 0.0015671561704948545,
+      "grad_norm": 0.0013897513272240758,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.0002,
+      "loss": 0.0007,
       "step": 7208
     },
     {
       "epoch": 68.0,
-      "eval_LOCATION_f1": 0.832214765100671,
+      "eval_LOCATION_f1": 0.7948717948717949,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7948717948717948,
+      "eval_LOCATION_precision": 0.7294117647058823,
       "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.8,
+      "eval_ORGANIZATION_f1": 0.7857142857142856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7466666666666667,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9096774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_PERSON_precision": 0.88125,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7868852459016394,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.7272727272727273,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.33385857939720154,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8591160220994475,
-      "eval_overall_precision": 0.8249336870026526,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9543,
-      "eval_samples_per_second": 195.947,
-      "eval_steps_per_second": 3.144,
+      "eval_loss": 0.3983727991580963,
+      "eval_overall_accuracy": 0.9597478176527643,
+      "eval_overall_f1": 0.841248303934871,
+      "eval_overall_precision": 0.7948717948717948,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3663,
+      "eval_samples_per_second": 510.562,
+      "eval_steps_per_second": 8.191,
       "step": 7208
     },
     {
       "epoch": 69.0,
-      "grad_norm": 0.0012689490104094148,
+      "grad_norm": 0.0465279296040535,
       "learning_rate": 1.55e-05,
-      "loss": 0.0006,
+      "loss": 0.0014,
       "step": 7314
     },
     {
       "epoch": 69.0,
-      "eval_LOCATION_f1": 0.8400000000000001,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_precision": 0.7682926829268293,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7215189873417722,
+      "eval_ORGANIZATION_f1": 0.7786259541984732,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6129032258064516,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9078947368421053,
+      "eval_ORGANIZATION_precision": 0.7727272727272727,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9161290322580645,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8961038961038961,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8875,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "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": 0.7999999999999999,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.35171428322792053,
-      "eval_overall_accuracy": 0.9580504364694471,
-      "eval_overall_f1": 0.8382749326145552,
-      "eval_overall_precision": 0.7873417721518987,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.953,
-      "eval_samples_per_second": 196.232,
-      "eval_steps_per_second": 3.148,
+      "eval_loss": 0.3407794237136841,
+      "eval_overall_accuracy": 0.9655674102812803,
+      "eval_overall_f1": 0.8547717842323651,
+      "eval_overall_precision": 0.8218085106382979,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3677,
+      "eval_samples_per_second": 508.566,
+      "eval_steps_per_second": 8.159,
       "step": 7314
     },
     {
       "epoch": 70.0,
-      "grad_norm": 0.00089789874618873,
+      "grad_norm": 0.0013118417700752616,
       "learning_rate": 1.5e-05,
-      "loss": 0.001,
+      "loss": 0.0003,
       "step": 7420
     },
     {
       "epoch": 70.0,
-      "eval_LOCATION_f1": 0.8378378378378378,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8051948051948052,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.802919708029197,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
+      "eval_ORGANIZATION_precision": 0.7638888888888888,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.32278671860694885,
-      "eval_overall_accuracy": 0.9648399612027158,
-      "eval_overall_f1": 0.853185595567867,
-      "eval_overall_precision": 0.8213333333333334,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 1.0201,
-      "eval_samples_per_second": 183.317,
-      "eval_steps_per_second": 2.941,
+      "eval_loss": 0.34899669885635376,
+      "eval_overall_accuracy": 0.9633850630455868,
+      "eval_overall_f1": 0.8555708390646491,
+      "eval_overall_precision": 0.8184210526315789,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3664,
+      "eval_samples_per_second": 510.394,
+      "eval_steps_per_second": 8.188,
       "step": 7420
     },
     {
       "epoch": 71.0,
-      "grad_norm": 0.001998705556616187,
+      "grad_norm": 0.0004932036972604692,
       "learning_rate": 1.45e-05,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 7526
     },
     {
       "epoch": 71.0,
-      "eval_LOCATION_f1": 0.8400000000000001,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_precision": 0.7682926829268293,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.8111888111888114,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7435897435897436,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9185667752442996,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8980891719745223,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3305073380470276,
-      "eval_overall_accuracy": 0.9633850630455868,
-      "eval_overall_f1": 0.8571428571428571,
-      "eval_overall_precision": 0.8188976377952756,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9607,
-      "eval_samples_per_second": 194.654,
-      "eval_steps_per_second": 3.123,
+      "eval_loss": 0.35279467701911926,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8512396694214875,
+      "eval_overall_precision": 0.8153034300791556,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3689,
+      "eval_samples_per_second": 506.928,
+      "eval_steps_per_second": 8.133,
       "step": 7526
     },
     {
       "epoch": 72.0,
-      "grad_norm": 0.0005069375038146973,
+      "grad_norm": 0.00035561618278734386,
       "learning_rate": 1.4000000000000001e-05,
-      "loss": 0.0006,
+      "loss": 0.0002,
       "step": 7632
     },
     {
       "epoch": 72.0,
-      "eval_LOCATION_f1": 0.8356164383561644,
+      "eval_LOCATION_f1": 0.8181818181818181,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.8133333333333334,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7826086956521738,
+      "eval_LOCATION_precision": 0.7590361445783133,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7883211678832116,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7397260273972602,
+      "eval_ORGANIZATION_precision": 0.75,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9271523178807947,
+      "eval_PERSON_f1": 0.9185667752442996,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9210526315789473,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8000000000000002,
+      "eval_PERSON_precision": 0.8980891719745223,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7567567567567568,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.30825239419937134,
-      "eval_overall_accuracy": 0.9675072744907857,
-      "eval_overall_f1": 0.8575418994413407,
-      "eval_overall_precision": 0.8319783197831978,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9597,
-      "eval_samples_per_second": 194.861,
-      "eval_steps_per_second": 3.126,
+      "eval_loss": 0.3692629337310791,
+      "eval_overall_accuracy": 0.965082444228904,
+      "eval_overall_f1": 0.8555708390646491,
+      "eval_overall_precision": 0.8184210526315789,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.3695,
+      "eval_samples_per_second": 506.096,
+      "eval_steps_per_second": 8.119,
       "step": 7632
     },
     {
       "epoch": 73.0,
-      "grad_norm": 0.0007079507340677083,
+      "grad_norm": 0.0001549980224808678,
       "learning_rate": 1.3500000000000001e-05,
-      "loss": 0.0003,
+      "loss": 0.0004,
       "step": 7738
     },
     {
       "epoch": 73.0,
-      "eval_LOCATION_f1": 0.8266666666666665,
+      "eval_LOCATION_f1": 0.8344370860927152,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7857142857142856,
+      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.8088235294117647,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_precision": 0.7746478873239436,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9205298013245033,
+      "eval_PERSON_f1": 0.9190938511326862,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9144736842105263,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8000000000000002,
+      "eval_PERSON_precision": 0.8930817610062893,
+      "eval_PERSON_recall": 0.9466666666666667,
+      "eval_QUANTITY_f1": 0.8285714285714285,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7567567567567568,
-      "eval_QUANTITY_recall": 0.8484848484848485,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_TIME_f1": 0.8275862068965518,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3165390193462372,
-      "eval_overall_accuracy": 0.9667798254122212,
-      "eval_overall_f1": 0.853185595567867,
-      "eval_overall_precision": 0.8213333333333334,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9558,
-      "eval_samples_per_second": 195.657,
-      "eval_steps_per_second": 3.139,
+      "eval_loss": 0.3441883325576782,
+      "eval_overall_accuracy": 0.966537342386033,
+      "eval_overall_f1": 0.8646408839779005,
+      "eval_overall_precision": 0.830238726790451,
+      "eval_overall_recall": 0.9020172910662824,
+      "eval_runtime": 0.3666,
+      "eval_samples_per_second": 510.05,
+      "eval_steps_per_second": 8.183,
       "step": 7738
     },
     {
       "epoch": 74.0,
-      "grad_norm": 0.00018394706421531737,
+      "grad_norm": 0.0007536042830906808,
       "learning_rate": 1.3000000000000001e-05,
-      "loss": 0.0002,
+      "loss": 0.0007,
       "step": 7844
     },
     {
       "epoch": 74.0,
-      "eval_LOCATION_f1": 0.8053691275167785,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.8085106382978723,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7851851851851851,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.75,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.9235880398671096,
+      "eval_ORGANIZATION_precision": 0.7571428571428571,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9205298013245033,
+      "eval_PERSON_precision": 0.879746835443038,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8000000000000002,
+      "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7567567567567568,
-      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_QUANTITY_precision": 0.7631578947368421,
+      "eval_QUANTITY_recall": 0.8787878787878788,
       "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.31818944215774536,
-      "eval_overall_accuracy": 0.967992240543162,
-      "eval_overall_f1": 0.8555555555555555,
-      "eval_overall_precision": 0.8257372654155496,
+      "eval_loss": 0.3682120442390442,
+      "eval_overall_accuracy": 0.9629000969932104,
+      "eval_overall_f1": 0.8461538461538461,
+      "eval_overall_precision": 0.8083989501312336,
       "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9609,
-      "eval_samples_per_second": 194.606,
-      "eval_steps_per_second": 3.122,
+      "eval_runtime": 0.3779,
+      "eval_samples_per_second": 494.866,
+      "eval_steps_per_second": 7.939,
       "step": 7844
     },
     {
       "epoch": 75.0,
-      "grad_norm": 0.00024878690601326525,
+      "grad_norm": 0.001529936445876956,
       "learning_rate": 1.25e-05,
-      "loss": 0.0012,
+      "loss": 0.0013,
       "step": 7950
     },
     {
       "epoch": 75.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.8082191780821918,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7883211678832116,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7283950617283951,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.75,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_precision": 0.8860759493670886,
       "eval_PERSON_recall": 0.9333333333333333,
       "eval_QUANTITY_f1": 0.8169014084507042,
       "eval_QUANTITY_number": 33,
       "eval_QUANTITY_precision": 0.7631578947368421,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.34358134865760803,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8528610354223433,
-      "eval_overall_precision": 0.8087855297157622,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.9518,
-      "eval_samples_per_second": 196.477,
-      "eval_steps_per_second": 3.152,
+      "eval_loss": 0.36499080061912537,
+      "eval_overall_accuracy": 0.9636275460717749,
+      "eval_overall_f1": 0.8543956043956045,
+      "eval_overall_precision": 0.8162729658792651,
+      "eval_overall_recall": 0.8962536023054755,
+      "eval_runtime": 0.365,
+      "eval_samples_per_second": 512.259,
+      "eval_steps_per_second": 8.218,
       "step": 7950
     },
     {
       "epoch": 76.0,
-      "grad_norm": 0.0002495927328709513,
+      "grad_norm": 0.00030084975878708065,
       "learning_rate": 1.2e-05,
-      "loss": 0.0002,
+      "loss": 0.0004,
       "step": 8056
     },
     {
       "epoch": 76.0,
-      "eval_LOCATION_f1": 0.8026315789473684,
+      "eval_LOCATION_f1": 0.8266666666666665,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.802721088435374,
+      "eval_LOCATION_precision": 0.7848101265822784,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7195121951219512,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9210526315789475,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9090909090909091,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_QUANTITY_f1": 0.8656716417910447,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
+      "eval_QUANTITY_precision": 0.8529411764705882,
       "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_TIME_f1": 0.8421052631578947,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.8275862068965517,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.34999027848243713,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.854024556616644,
-      "eval_overall_precision": 0.810880829015544,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.9639,
-      "eval_samples_per_second": 194.004,
-      "eval_steps_per_second": 3.112,
+      "eval_loss": 0.3499511778354645,
+      "eval_overall_accuracy": 0.9655674102812803,
+      "eval_overall_f1": 0.8547717842323651,
+      "eval_overall_precision": 0.8218085106382979,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3692,
+      "eval_samples_per_second": 506.48,
+      "eval_steps_per_second": 8.125,
       "step": 8056
     },
     {
       "epoch": 77.0,
-      "grad_norm": 0.000243131464230828,
+      "grad_norm": 0.0008816963527351618,
       "learning_rate": 1.1500000000000002e-05,
-      "loss": 0.0007,
+      "loss": 0.0008,
       "step": 8162
     },
     {
       "epoch": 77.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8076923076923077,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.8137931034482759,
+      "eval_LOCATION_precision": 0.7411764705882353,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.759124087591241,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7375,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9210526315789475,
+      "eval_ORGANIZATION_precision": 0.7222222222222222,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9090909090909091,
+      "eval_PERSON_precision": 0.8860759493670886,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7868852459016394,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.75,
+      "eval_TIME_precision": 0.7272727272727273,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3344757556915283,
-      "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8575342465753425,
-      "eval_overall_precision": 0.8172323759791122,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.9601,
-      "eval_samples_per_second": 194.77,
-      "eval_steps_per_second": 3.125,
+      "eval_loss": 0.36515700817108154,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8376534788540247,
+      "eval_overall_precision": 0.7953367875647669,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3678,
+      "eval_samples_per_second": 508.497,
+      "eval_steps_per_second": 8.158,
       "step": 8162
     },
     {
       "epoch": 78.0,
-      "grad_norm": 0.00021198501053731889,
+      "grad_norm": 0.007112889084964991,
       "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.0006,
+      "loss": 0.0002,
       "step": 8268
     },
     {
       "epoch": 78.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8181818181818181,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.802919708029197,
+      "eval_LOCATION_precision": 0.7590361445783133,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7638888888888888,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9210526315789475,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9090909090909091,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3224840760231018,
-      "eval_overall_accuracy": 0.9655674102812803,
-      "eval_overall_f1": 0.8559556786703602,
-      "eval_overall_precision": 0.824,
-      "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 0.9532,
-      "eval_samples_per_second": 196.177,
-      "eval_steps_per_second": 3.147,
+      "eval_loss": 0.3537391126155853,
+      "eval_overall_accuracy": 0.9638700290979632,
+      "eval_overall_f1": 0.8426812585499316,
+      "eval_overall_precision": 0.8020833333333334,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3745,
+      "eval_samples_per_second": 499.333,
+      "eval_steps_per_second": 8.011,
       "step": 8268
     },
     {
       "epoch": 79.0,
-      "grad_norm": 0.0016178454970940948,
+      "grad_norm": 0.0005999546265229583,
       "learning_rate": 1.05e-05,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 8374
     },
     {
       "epoch": 79.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "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": 0.8275862068965518,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3340538740158081,
+      "eval_loss": 0.35688579082489014,
       "eval_overall_accuracy": 0.965082444228904,
-      "eval_overall_f1": 0.8547717842323651,
-      "eval_overall_precision": 0.8218085106382979,
+      "eval_overall_f1": 0.8477366255144033,
+      "eval_overall_precision": 0.8089005235602095,
       "eval_overall_recall": 0.8904899135446686,
-      "eval_runtime": 1.013,
-      "eval_samples_per_second": 184.602,
-      "eval_steps_per_second": 2.962,
+      "eval_runtime": 0.3639,
+      "eval_samples_per_second": 513.835,
+      "eval_steps_per_second": 8.243,
       "step": 8374
     },
     {
       "epoch": 80.0,
-      "grad_norm": 0.0010399464517831802,
+      "grad_norm": 0.0004196091031190008,
       "learning_rate": 1e-05,
       "loss": 0.0005,
       "step": 8480
     },
     {
       "epoch": 80.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7913669064748202,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7851851851851851,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7432432432432432,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_ORGANIZATION_precision": 0.7571428571428571,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3451690971851349,
-      "eval_overall_accuracy": 0.965082444228904,
-      "eval_overall_f1": 0.8543689320388349,
-      "eval_overall_precision": 0.8235294117647058,
+      "eval_loss": 0.35123497247695923,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8473177441540578,
+      "eval_overall_precision": 0.8105263157894737,
       "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9594,
-      "eval_samples_per_second": 194.923,
-      "eval_steps_per_second": 3.127,
+      "eval_runtime": 0.37,
+      "eval_samples_per_second": 505.398,
+      "eval_steps_per_second": 8.108,
       "step": 8480
     },
     {
       "epoch": 81.0,
-      "grad_norm": 0.0004898614133708179,
+      "grad_norm": 0.0005896950024180114,
       "learning_rate": 9.5e-06,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 8586
     },
     {
       "epoch": 81.0,
-      "eval_LOCATION_f1": 0.8133333333333335,
+      "eval_LOCATION_f1": 0.8181818181818181,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7857142857142856,
+      "eval_LOCATION_precision": 0.7590361445783133,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7913669064748202,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_precision": 0.7432432432432432,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9144736842105262,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9025974025974026,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8275862068965518,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.8135593220338982,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.8,
+      "eval_TIME_precision": 0.7741935483870968,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3453369140625,
-      "eval_overall_accuracy": 0.965082444228904,
-      "eval_overall_f1": 0.853185595567867,
-      "eval_overall_precision": 0.8213333333333334,
-      "eval_overall_recall": 0.8876080691642652,
-      "eval_runtime": 0.9528,
-      "eval_samples_per_second": 196.26,
-      "eval_steps_per_second": 3.149,
+      "eval_loss": 0.35577791929244995,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3645,
+      "eval_samples_per_second": 513.064,
+      "eval_steps_per_second": 8.231,
       "step": 8586
     },
     {
       "epoch": 82.0,
-      "grad_norm": 0.0012529068626463413,
+      "grad_norm": 0.0008361159707419574,
       "learning_rate": 9e-06,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 8692
     },
     {
       "epoch": 82.0,
-      "eval_LOCATION_f1": 0.8076923076923077,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7411764705882353,
+      "eval_LOCATION_precision": 0.7682926829268293,
       "eval_LOCATION_recall": 0.8873239436619719,
-      "eval_ORGANIZATION_f1": 0.7943262411347517,
+      "eval_ORGANIZATION_f1": 0.787878787878788,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7368421052631579,
-      "eval_ORGANIZATION_recall": 0.8615384615384616,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_ORGANIZATION_precision": 0.7761194029850746,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "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": 0.3554005026817322,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8481532147742818,
-      "eval_overall_precision": 0.8072916666666666,
-      "eval_overall_recall": 0.8933717579250721,
-      "eval_runtime": 0.9565,
-      "eval_samples_per_second": 195.504,
-      "eval_steps_per_second": 3.136,
+      "eval_loss": 0.3439392149448395,
+      "eval_overall_accuracy": 0.9660523763336566,
+      "eval_overall_f1": 0.8539638386648122,
+      "eval_overall_precision": 0.8252688172043011,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3724,
+      "eval_samples_per_second": 502.206,
+      "eval_steps_per_second": 8.057,
       "step": 8692
     },
     {
       "epoch": 83.0,
-      "grad_norm": 0.0037664775736629963,
+      "grad_norm": 0.0006481435848399997,
       "learning_rate": 8.500000000000002e-06,
-      "loss": 0.0005,
+      "loss": 0.0002,
       "step": 8798
     },
     {
       "epoch": 83.0,
-      "eval_LOCATION_f1": 0.8079470198675496,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7681159420289856,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7819548872180452,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.726027397260274,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.9102990033222591,
+      "eval_ORGANIZATION_precision": 0.7647058823529411,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.9150326797385621,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9072847682119205,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.8974358974358975,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "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": 0.33550789952278137,
+      "eval_loss": 0.35054975748062134,
       "eval_overall_accuracy": 0.965082444228904,
-      "eval_overall_f1": 0.8491620111731844,
-      "eval_overall_precision": 0.8238482384823849,
-      "eval_overall_recall": 0.8760806916426513,
-      "eval_runtime": 0.9544,
-      "eval_samples_per_second": 195.933,
-      "eval_steps_per_second": 3.143,
+      "eval_overall_f1": 0.8527777777777777,
+      "eval_overall_precision": 0.8230563002680965,
+      "eval_overall_recall": 0.8847262247838616,
+      "eval_runtime": 0.3688,
+      "eval_samples_per_second": 507.07,
+      "eval_steps_per_second": 8.135,
       "step": 8798
     },
     {
       "epoch": 84.0,
-      "grad_norm": 0.00019280197739135474,
+      "grad_norm": 0.001612077234312892,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.0002,
+      "loss": 0.0006,
       "step": 8904
     },
     {
       "epoch": 84.0,
-      "eval_LOCATION_f1": 0.7922077922077921,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7659574468085107,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7826086956521738,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7105263157894737,
+      "eval_ORGANIZATION_precision": 0.7397260273972602,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7868852459016394,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.7272727272727273,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3492893576622009,
-      "eval_overall_accuracy": 0.9633850630455868,
-      "eval_overall_f1": 0.8422496570644719,
-      "eval_overall_precision": 0.8036649214659686,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9553,
-      "eval_samples_per_second": 195.753,
-      "eval_steps_per_second": 3.14,
+      "eval_loss": 0.3632793724536896,
+      "eval_overall_accuracy": 0.9631425800193987,
+      "eval_overall_f1": 0.8481532147742818,
+      "eval_overall_precision": 0.8072916666666666,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.371,
+      "eval_samples_per_second": 504.03,
+      "eval_steps_per_second": 8.086,
       "step": 8904
     },
     {
       "epoch": 85.0,
-      "grad_norm": 0.0007057705661281943,
+      "grad_norm": 0.001582043943926692,
       "learning_rate": 7.5e-06,
       "loss": 0.0003,
       "step": 9010
     },
     {
       "epoch": 85.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7714285714285715,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7941176470588235,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.72,
+      "eval_ORGANIZATION_precision": 0.7605633802816901,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3493691086769104,
-      "eval_overall_accuracy": 0.9636275460717749,
-      "eval_overall_f1": 0.8418156808803301,
-      "eval_overall_precision": 0.8052631578947368,
-      "eval_overall_recall": 0.8818443804034583,
-      "eval_runtime": 0.9473,
-      "eval_samples_per_second": 197.399,
-      "eval_steps_per_second": 3.167,
+      "eval_loss": 0.36183223128318787,
+      "eval_overall_accuracy": 0.9641125121241513,
+      "eval_overall_f1": 0.8500687757909215,
+      "eval_overall_precision": 0.8131578947368421,
+      "eval_overall_recall": 0.8904899135446686,
+      "eval_runtime": 0.3683,
+      "eval_samples_per_second": 507.773,
+      "eval_steps_per_second": 8.146,
       "step": 9010
     },
     {
       "epoch": 86.0,
-      "grad_norm": 0.00023777979367878288,
+      "grad_norm": 0.0002886342990677804,
       "learning_rate": 7.000000000000001e-06,
       "loss": 0.0002,
       "step": 9116
     },
     {
       "epoch": 86.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.8421052631578947,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7801418439716312,
+      "eval_LOCATION_precision": 0.7901234567901234,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7883211678832116,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7236842105263158,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.9049180327868852,
+      "eval_ORGANIZATION_precision": 0.75,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8903225806451613,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.35241079330444336,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8434065934065933,
-      "eval_overall_precision": 0.8057742782152231,
-      "eval_overall_recall": 0.8847262247838616,
-      "eval_runtime": 0.9506,
-      "eval_samples_per_second": 196.717,
-      "eval_steps_per_second": 3.156,
+      "eval_loss": 0.3576675355434418,
+      "eval_overall_accuracy": 0.9643549951503395,
+      "eval_overall_f1": 0.8528198074277854,
+      "eval_overall_precision": 0.8157894736842105,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3791,
+      "eval_samples_per_second": 493.217,
+      "eval_steps_per_second": 7.913,
       "step": 9116
     },
     {
       "epoch": 87.0,
-      "grad_norm": 0.0003070076636504382,
+      "grad_norm": 0.00032167613971978426,
       "learning_rate": 6.5000000000000004e-06,
       "loss": 0.0002,
       "step": 9222
     },
     {
       "epoch": 87.0,
-      "eval_LOCATION_f1": 0.8104575163398693,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7560975609756098,
-      "eval_LOCATION_recall": 0.8732394366197183,
-      "eval_ORGANIZATION_f1": 0.7785234899328859,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7769784172661871,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6904761904761905,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.7297297297297297,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8285714285714285,
+      "eval_PERSON_precision": 0.89171974522293,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7837837837837838,
-      "eval_QUANTITY_recall": 0.8787878787878788,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3664219081401825,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8466757123473542,
-      "eval_overall_precision": 0.8,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9567,
-      "eval_samples_per_second": 195.468,
-      "eval_steps_per_second": 3.136,
+      "eval_loss": 0.3611268699169159,
+      "eval_overall_accuracy": 0.9636275460717749,
+      "eval_overall_f1": 0.8493150684931507,
+      "eval_overall_precision": 0.8093994778067886,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3664,
+      "eval_samples_per_second": 510.384,
+      "eval_steps_per_second": 8.188,
       "step": 9222
     },
     {
       "epoch": 88.0,
-      "grad_norm": 0.0003128907410427928,
+      "grad_norm": 0.0001423046487616375,
       "learning_rate": 6e-06,
       "loss": 0.0002,
       "step": 9328
     },
     {
       "epoch": 88.0,
-      "eval_LOCATION_f1": 0.8053691275167785,
+      "eval_LOCATION_f1": 0.823529411764706,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7792207792207791,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7969924812030074,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6741573033707865,
-      "eval_ORGANIZATION_recall": 0.9230769230769231,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7794117647058824,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8450704225352113,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.7894736842105263,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3721804618835449,
-      "eval_overall_accuracy": 0.9626576139670223,
-      "eval_overall_f1": 0.8509485094850948,
-      "eval_overall_precision": 0.80306905370844,
-      "eval_overall_recall": 0.9048991354466859,
-      "eval_runtime": 0.9518,
-      "eval_samples_per_second": 196.461,
-      "eval_steps_per_second": 3.152,
+      "eval_loss": 0.3588053584098816,
+      "eval_overall_accuracy": 0.9655674102812803,
+      "eval_overall_f1": 0.8520055325034579,
+      "eval_overall_precision": 0.8191489361702128,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3692,
+      "eval_samples_per_second": 506.568,
+      "eval_steps_per_second": 8.127,
       "step": 9328
     },
     {
       "epoch": 89.0,
-      "grad_norm": 0.0002441601827740669,
+      "grad_norm": 0.00040084519423544407,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 9434
     },
     {
       "epoch": 89.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8344370860927152,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7837837837837838,
+      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6987951807228916,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3669990599155426,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8501362397820164,
-      "eval_overall_precision": 0.8062015503875969,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.968,
-      "eval_samples_per_second": 193.188,
-      "eval_steps_per_second": 3.099,
+      "eval_loss": 0.3615128695964813,
+      "eval_overall_accuracy": 0.9648399612027158,
+      "eval_overall_f1": 0.8496551724137931,
+      "eval_overall_precision": 0.8148148148148148,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3697,
+      "eval_samples_per_second": 505.85,
+      "eval_steps_per_second": 8.115,
       "step": 9434
     },
     {
       "epoch": 90.0,
-      "grad_norm": 0.001270479173399508,
+      "grad_norm": 0.0024764875415712595,
       "learning_rate": 5e-06,
-      "loss": 0.0002,
+      "loss": 0.0003,
       "step": 9540
     },
     {
       "epoch": 90.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8289473684210527,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7837837837837838,
+      "eval_LOCATION_precision": 0.7777777777777778,
+      "eval_LOCATION_recall": 0.8873239436619719,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6987951807228916,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9120521172638437,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
+      "eval_PERSON_precision": 0.89171974522293,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.36801740527153015,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8501362397820164,
-      "eval_overall_precision": 0.8062015503875969,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9464,
-      "eval_samples_per_second": 197.59,
-      "eval_steps_per_second": 3.17,
+      "eval_loss": 0.36288607120513916,
+      "eval_overall_accuracy": 0.9643549951503395,
+      "eval_overall_f1": 0.8496551724137931,
+      "eval_overall_precision": 0.8148148148148148,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3743,
+      "eval_samples_per_second": 499.595,
+      "eval_steps_per_second": 8.015,
       "step": 9540
     },
     {
       "epoch": 91.0,
-      "grad_norm": 0.00022247471497394145,
+      "grad_norm": 0.000249840464675799,
       "learning_rate": 4.5e-06,
       "loss": 0.0002,
       "step": 9646
     },
     {
       "epoch": 91.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7837837837837838,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6987951807228916,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9084967320261438,
-      "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
+      "eval_PERSON_number": 150,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3627462089061737,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8462585034013606,
-      "eval_overall_precision": 0.8015463917525774,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9545,
-      "eval_samples_per_second": 195.918,
-      "eval_steps_per_second": 3.143,
+      "eval_loss": 0.3642159104347229,
+      "eval_overall_accuracy": 0.9648399612027158,
+      "eval_overall_f1": 0.8528198074277854,
+      "eval_overall_precision": 0.8157894736842105,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3698,
+      "eval_samples_per_second": 505.704,
+      "eval_steps_per_second": 8.113,
       "step": 9646
     },
     {
       "epoch": 92.0,
-      "grad_norm": 0.00013813188706990331,
+      "grad_norm": 0.00021500952425412834,
       "learning_rate": 4.000000000000001e-06,
       "loss": 0.0002,
       "step": 9752
     },
     {
       "epoch": 92.0,
-      "eval_LOCATION_f1": 0.7973856209150327,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8591549295774648,
-      "eval_ORGANIZATION_f1": 0.7785234899328859,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6904761904761905,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3727177083492279,
-      "eval_overall_accuracy": 0.9629000969932104,
-      "eval_overall_f1": 0.8493894165535958,
-      "eval_overall_precision": 0.8025641025641026,
-      "eval_overall_recall": 0.9020172910662824,
-      "eval_runtime": 0.954,
-      "eval_samples_per_second": 196.013,
-      "eval_steps_per_second": 3.145,
+      "eval_loss": 0.3649475574493408,
+      "eval_overall_accuracy": 0.9648399612027158,
+      "eval_overall_f1": 0.8528198074277854,
+      "eval_overall_precision": 0.8157894736842105,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3773,
+      "eval_samples_per_second": 495.618,
+      "eval_steps_per_second": 7.951,
       "step": 9752
     },
     {
       "epoch": 93.0,
-      "grad_norm": 0.000215419742744416,
+      "grad_norm": 0.000220635934965685,
       "learning_rate": 3.5000000000000004e-06,
       "loss": 0.0002,
       "step": 9858
     },
     {
       "epoch": 93.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8157894736842104,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7837837837837838,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6987951807228916,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9084967320261438,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8910256410256411,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
       "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
       "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.37157806754112244,
-      "eval_overall_accuracy": 0.962172647914646,
-      "eval_overall_f1": 0.8451086956521738,
-      "eval_overall_precision": 0.7994858611825193,
-      "eval_overall_recall": 0.8962536023054755,
-      "eval_runtime": 0.9451,
-      "eval_samples_per_second": 197.862,
-      "eval_steps_per_second": 3.174,
+      "eval_loss": 0.36348941922187805,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.8484848484848485,
+      "eval_overall_precision": 0.8126649076517151,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3679,
+      "eval_samples_per_second": 508.312,
+      "eval_steps_per_second": 8.155,
       "step": 9858
     },
     {
       "epoch": 94.0,
-      "grad_norm": 0.0003316714719403535,
+      "grad_norm": 0.00017434197070542723,
       "learning_rate": 3e-06,
       "loss": 0.0002,
       "step": 9964
     },
     {
       "epoch": 94.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8157894736842104,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.8082191780821918,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8732394366197183,
+      "eval_ORGANIZATION_f1": 0.7794117647058824,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7283950617283951,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9205298013245033,
+      "eval_ORGANIZATION_precision": 0.7464788732394366,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9144736842105263,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.8000000000000002,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7567567567567568,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.35780400037765503,
+      "eval_loss": 0.3645237386226654,
       "eval_overall_accuracy": 0.9643549951503395,
-      "eval_overall_f1": 0.8559670781893004,
-      "eval_overall_precision": 0.8167539267015707,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9435,
-      "eval_samples_per_second": 198.203,
-      "eval_steps_per_second": 3.18,
+      "eval_overall_f1": 0.8484848484848485,
+      "eval_overall_precision": 0.8126649076517151,
+      "eval_overall_recall": 0.8876080691642652,
+      "eval_runtime": 0.3728,
+      "eval_samples_per_second": 501.672,
+      "eval_steps_per_second": 8.048,
       "step": 9964
     },
     {
       "epoch": 95.0,
-      "grad_norm": 0.005569119472056627,
+      "grad_norm": 0.0002888070303015411,
       "learning_rate": 2.5e-06,
-      "loss": 0.0002,
+      "loss": 0.0005,
       "step": 10070
     },
     {
       "epoch": 95.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.8082191780821918,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7283950617283951,
-      "eval_ORGANIZATION_recall": 0.9076923076923077,
-      "eval_PERSON_f1": 0.9205298013245033,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9144736842105263,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3565683662891388,
+      "eval_loss": 0.3668120801448822,
       "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8559670781893004,
-      "eval_overall_precision": 0.8167539267015707,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9611,
-      "eval_samples_per_second": 194.577,
-      "eval_steps_per_second": 3.122,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3658,
+      "eval_samples_per_second": 511.262,
+      "eval_steps_per_second": 8.202,
       "step": 10070
     },
     {
       "epoch": 96.0,
-      "grad_norm": 0.00045707172830589116,
+      "grad_norm": 0.0007120606023818254,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.0006,
+      "loss": 0.0002,
       "step": 10176
     },
     {
       "epoch": 96.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3609143793582916,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8524590163934427,
-      "eval_overall_precision": 0.8103896103896104,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9457,
-      "eval_samples_per_second": 197.739,
-      "eval_steps_per_second": 3.172,
+      "eval_loss": 0.36810311675071716,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3867,
+      "eval_samples_per_second": 483.598,
+      "eval_steps_per_second": 7.758,
       "step": 10176
     },
     {
       "epoch": 97.0,
-      "grad_norm": 0.0002619586593937129,
+      "grad_norm": 0.00018466792243998498,
       "learning_rate": 1.5e-06,
-      "loss": 0.0003,
+      "loss": 0.0002,
       "step": 10282
     },
     {
       "epoch": 97.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3629205524921417,
-      "eval_overall_accuracy": 0.9641125121241513,
-      "eval_overall_f1": 0.8524590163934427,
-      "eval_overall_precision": 0.8103896103896104,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9477,
-      "eval_samples_per_second": 197.31,
-      "eval_steps_per_second": 3.165,
+      "eval_loss": 0.36953815817832947,
+      "eval_overall_accuracy": 0.9645974781765276,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3666,
+      "eval_samples_per_second": 510.133,
+      "eval_steps_per_second": 8.184,
       "step": 10282
     },
     {
       "epoch": 98.0,
-      "grad_norm": 0.0001948641292983666,
+      "grad_norm": 0.008206859230995178,
       "learning_rate": 1.0000000000000002e-06,
       "loss": 0.0002,
       "step": 10388
     },
     {
       "epoch": 98.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.7804878048780488,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.35546010732650757,
+      "eval_loss": 0.36951956152915955,
       "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8524590163934427,
-      "eval_overall_precision": 0.8103896103896104,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9637,
-      "eval_samples_per_second": 194.037,
-      "eval_steps_per_second": 3.113,
+      "eval_overall_f1": 0.850480109739369,
+      "eval_overall_precision": 0.8115183246073299,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3722,
+      "eval_samples_per_second": 502.461,
+      "eval_steps_per_second": 8.061,
       "step": 10388
     },
     {
       "epoch": 99.0,
-      "grad_norm": 0.023726776242256165,
+      "grad_norm": 0.00013635109644383192,
       "learning_rate": 5.000000000000001e-07,
       "loss": 0.0002,
       "step": 10494
     },
     {
       "epoch": 99.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.35447004437446594,
-      "eval_overall_accuracy": 0.9648399612027158,
-      "eval_overall_f1": 0.8524590163934427,
-      "eval_overall_precision": 0.8103896103896104,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.951,
-      "eval_samples_per_second": 196.644,
-      "eval_steps_per_second": 3.155,
+      "eval_loss": 0.3702433109283447,
+      "eval_overall_accuracy": 0.9643549951503395,
+      "eval_overall_f1": 0.8493150684931507,
+      "eval_overall_precision": 0.8093994778067886,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3665,
+      "eval_samples_per_second": 510.199,
+      "eval_steps_per_second": 8.185,
       "step": 10494
     },
     {
       "epoch": 100.0,
-      "grad_norm": 0.0003471510426606983,
+      "grad_norm": 0.00016493220755364746,
       "learning_rate": 0.0,
-      "loss": 0.0004,
+      "loss": 0.0002,
       "step": 10600
     },
     {
       "epoch": 100.0,
-      "eval_LOCATION_f1": 0.7894736842105262,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 71,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.8450704225352113,
-      "eval_ORGANIZATION_f1": 0.7945205479452054,
+      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_recall": 0.9014084507042254,
+      "eval_ORGANIZATION_f1": 0.7737226277372263,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7160493827160493,
-      "eval_ORGANIZATION_recall": 0.8923076923076924,
-      "eval_PERSON_f1": 0.9180327868852459,
+      "eval_ORGANIZATION_precision": 0.7361111111111112,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9155844155844156,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.9032258064516129,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.8571428571428571,
+      "eval_PERSON_precision": 0.8924050632911392,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7887323943661972,
       "eval_QUANTITY_number": 33,
-      "eval_QUANTITY_precision": 0.8108108108108109,
-      "eval_QUANTITY_recall": 0.9090909090909091,
-      "eval_TIME_f1": 0.8135593220338982,
+      "eval_QUANTITY_precision": 0.7368421052631579,
+      "eval_QUANTITY_recall": 0.8484848484848485,
+      "eval_TIME_f1": 0.7999999999999999,
       "eval_TIME_number": 28,
-      "eval_TIME_precision": 0.7741935483870968,
+      "eval_TIME_precision": 0.75,
       "eval_TIME_recall": 0.8571428571428571,
-      "eval_loss": 0.3544124364852905,
-      "eval_overall_accuracy": 0.9645974781765276,
-      "eval_overall_f1": 0.8524590163934427,
-      "eval_overall_precision": 0.8103896103896104,
-      "eval_overall_recall": 0.899135446685879,
-      "eval_runtime": 0.9623,
-      "eval_samples_per_second": 194.318,
-      "eval_steps_per_second": 3.117,
+      "eval_loss": 0.370334655046463,
+      "eval_overall_accuracy": 0.9643549951503395,
+      "eval_overall_f1": 0.8493150684931507,
+      "eval_overall_precision": 0.8093994778067886,
+      "eval_overall_recall": 0.8933717579250721,
+      "eval_runtime": 0.3682,
+      "eval_samples_per_second": 507.932,
+      "eval_steps_per_second": 8.149,
       "step": 10600
     },
     {
       "epoch": 100.0,
       "step": 10600,
       "total_flos": 4591573981087056.0,
-      "train_loss": 0.008363780669411118,
-      "train_runtime": 2646.2436,
-      "train_samples_per_second": 63.789,
-      "train_steps_per_second": 4.006
+      "train_loss": 0.00907013229428316,
+      "train_runtime": 992.7209,
+      "train_samples_per_second": 170.038,
+      "train_steps_per_second": 10.678
     }
   ],
   "logging_steps": 500,