diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -10,9 +10,9 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 1.6121073961257935,
+      "grad_norm": 1.5585302114486694,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 1.0203,
+      "loss": 1.0106,
       "step": 106
     },
     {
@@ -25,345 +25,345 @@
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.0,
       "eval_ORGANIZATION_recall": 0.0,
-      "eval_PERSON_f1": 0.03508771929824562,
+      "eval_PERSON_f1": 0.02366863905325444,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.14285714285714285,
-      "eval_PERSON_recall": 0.02,
+      "eval_PERSON_precision": 0.10526315789473684,
+      "eval_PERSON_recall": 0.013333333333333334,
       "eval_QUANTITY_f1": 0.0,
       "eval_QUANTITY_number": 29,
       "eval_QUANTITY_precision": 0.0,
       "eval_QUANTITY_recall": 0.0,
-      "eval_TIME_f1": 0.05405405405405406,
-      "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.3333333333333333,
-      "eval_TIME_recall": 0.029411764705882353,
-      "eval_loss": 0.591905415058136,
-      "eval_overall_accuracy": 0.840019689884322,
-      "eval_overall_f1": 0.020618556701030927,
-      "eval_overall_precision": 0.10810810810810811,
-      "eval_overall_recall": 0.011396011396011397,
-      "eval_runtime": 0.9779,
-      "eval_samples_per_second": 191.225,
-      "eval_steps_per_second": 3.068,
+      "eval_TIME_f1": 0.0,
+      "eval_TIME_number": 34,
+      "eval_TIME_precision": 0.0,
+      "eval_TIME_recall": 0.0,
+      "eval_loss": 0.584353506565094,
+      "eval_overall_accuracy": 0.840265813438346,
+      "eval_overall_f1": 0.010416666666666668,
+      "eval_overall_precision": 0.06060606060606061,
+      "eval_overall_recall": 0.005698005698005698,
+      "eval_runtime": 0.398,
+      "eval_samples_per_second": 469.835,
+      "eval_steps_per_second": 7.537,
       "step": 106
     },
     {
       "epoch": 2.0,
-      "grad_norm": 1.089346170425415,
+      "grad_norm": 1.2475420236587524,
       "learning_rate": 4.9e-05,
-      "loss": 0.5053,
+      "loss": 0.4994,
       "step": 212
     },
     {
       "epoch": 2.0,
-      "eval_LOCATION_f1": 0.2777777777777778,
-      "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.42857142857142855,
-      "eval_LOCATION_recall": 0.2054794520547945,
-      "eval_ORGANIZATION_f1": 0.06557377049180328,
-      "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.07017543859649122,
-      "eval_ORGANIZATION_recall": 0.06153846153846154,
-      "eval_PERSON_f1": 0.6512968299711815,
-      "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.5736040609137056,
-      "eval_PERSON_recall": 0.7533333333333333,
-      "eval_QUANTITY_f1": 0.08888888888888889,
-      "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.125,
-      "eval_QUANTITY_recall": 0.06896551724137931,
-      "eval_TIME_f1": 0.5074626865671642,
-      "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.5151515151515151,
-      "eval_TIME_recall": 0.5,
-      "eval_loss": 0.3727489709854126,
-      "eval_overall_accuracy": 0.8951513659857249,
-      "eval_overall_f1": 0.4383164005805515,
-      "eval_overall_precision": 0.4467455621301775,
+      "eval_LOCATION_f1": 0.2476190476190476,
+      "eval_LOCATION_number": 73,
+      "eval_LOCATION_precision": 0.40625,
+      "eval_LOCATION_recall": 0.1780821917808219,
+      "eval_ORGANIZATION_f1": 0.05309734513274336,
+      "eval_ORGANIZATION_number": 65,
+      "eval_ORGANIZATION_precision": 0.0625,
+      "eval_ORGANIZATION_recall": 0.046153846153846156,
+      "eval_PERSON_f1": 0.6495726495726495,
+      "eval_PERSON_number": 150,
+      "eval_PERSON_precision": 0.5671641791044776,
+      "eval_PERSON_recall": 0.76,
+      "eval_QUANTITY_f1": 0.11764705882352941,
+      "eval_QUANTITY_number": 29,
+      "eval_QUANTITY_precision": 0.13636363636363635,
+      "eval_QUANTITY_recall": 0.10344827586206896,
+      "eval_TIME_f1": 0.5454545454545455,
+      "eval_TIME_number": 34,
+      "eval_TIME_precision": 0.5625,
+      "eval_TIME_recall": 0.5294117647058824,
+      "eval_loss": 0.3691530227661133,
+      "eval_overall_accuracy": 0.8973664779719419,
+      "eval_overall_f1": 0.4402332361516035,
+      "eval_overall_precision": 0.4507462686567164,
       "eval_overall_recall": 0.4301994301994302,
-      "eval_runtime": 1.0224,
-      "eval_samples_per_second": 182.9,
-      "eval_steps_per_second": 2.934,
+      "eval_runtime": 0.386,
+      "eval_samples_per_second": 484.468,
+      "eval_steps_per_second": 7.772,
       "step": 212
     },
     {
       "epoch": 3.0,
-      "grad_norm": 0.9725346565246582,
+      "grad_norm": 1.0947109460830688,
       "learning_rate": 4.85e-05,
-      "loss": 0.3221,
+      "loss": 0.3195,
       "step": 318
     },
     {
       "epoch": 3.0,
-      "eval_LOCATION_f1": 0.6227544910179641,
+      "eval_LOCATION_f1": 0.6257668711656442,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.5531914893617021,
-      "eval_LOCATION_recall": 0.7123287671232876,
-      "eval_ORGANIZATION_f1": 0.2689075630252101,
+      "eval_LOCATION_precision": 0.5666666666666667,
+      "eval_LOCATION_recall": 0.6986301369863014,
+      "eval_ORGANIZATION_f1": 0.3464566929133858,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.2962962962962963,
-      "eval_ORGANIZATION_recall": 0.24615384615384617,
-      "eval_PERSON_f1": 0.8145896656534953,
+      "eval_ORGANIZATION_precision": 0.3548387096774194,
+      "eval_ORGANIZATION_recall": 0.3384615384615385,
+      "eval_PERSON_f1": 0.8220858895705522,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.7486033519553073,
+      "eval_PERSON_precision": 0.7613636363636364,
       "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.5588235294117647,
+      "eval_QUANTITY_f1": 0.5396825396825397,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.48717948717948717,
-      "eval_QUANTITY_recall": 0.6551724137931034,
-      "eval_TIME_f1": 0.8169014084507041,
+      "eval_QUANTITY_precision": 0.5,
+      "eval_QUANTITY_recall": 0.5862068965517241,
+      "eval_TIME_f1": 0.8285714285714286,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7837837837837838,
+      "eval_TIME_precision": 0.8055555555555556,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.21981501579284668,
-      "eval_overall_accuracy": 0.9305931577651981,
-      "eval_overall_f1": 0.6631299734748011,
-      "eval_overall_precision": 0.6203473945409429,
-      "eval_overall_recall": 0.7122507122507122,
-      "eval_runtime": 1.0345,
-      "eval_samples_per_second": 180.757,
-      "eval_steps_per_second": 2.9,
+      "eval_loss": 0.2105339616537094,
+      "eval_overall_accuracy": 0.9308392813192222,
+      "eval_overall_f1": 0.6755674232309746,
+      "eval_overall_precision": 0.635678391959799,
+      "eval_overall_recall": 0.7207977207977208,
+      "eval_runtime": 0.3847,
+      "eval_samples_per_second": 486.037,
+      "eval_steps_per_second": 7.797,
       "step": 318
     },
     {
       "epoch": 4.0,
-      "grad_norm": 0.8557797074317932,
+      "grad_norm": 1.0842955112457275,
       "learning_rate": 4.8e-05,
-      "loss": 0.2222,
+      "loss": 0.2206,
       "step": 424
     },
     {
       "epoch": 4.0,
-      "eval_LOCATION_f1": 0.6709677419354839,
+      "eval_LOCATION_f1": 0.7204968944099379,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6341463414634146,
-      "eval_LOCATION_recall": 0.7123287671232876,
-      "eval_ORGANIZATION_f1": 0.5844155844155845,
+      "eval_LOCATION_precision": 0.6590909090909091,
+      "eval_LOCATION_recall": 0.7945205479452054,
+      "eval_ORGANIZATION_f1": 0.5889570552147239,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5056179775280899,
-      "eval_ORGANIZATION_recall": 0.6923076923076923,
-      "eval_PERSON_f1": 0.8411214953271027,
+      "eval_ORGANIZATION_precision": 0.4897959183673469,
+      "eval_ORGANIZATION_recall": 0.7384615384615385,
+      "eval_PERSON_f1": 0.8553459119496856,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.7894736842105263,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.6567164179104478,
+      "eval_PERSON_precision": 0.8095238095238095,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.5753424657534246,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.5789473684210527,
-      "eval_QUANTITY_recall": 0.7586206896551724,
-      "eval_TIME_f1": 0.8219178082191781,
-      "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7692307692307693,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17654158174991608,
-      "eval_overall_accuracy": 0.9416687176962836,
-      "eval_overall_f1": 0.7376623376623377,
-      "eval_overall_precision": 0.6778042959427207,
-      "eval_overall_recall": 0.8091168091168092,
-      "eval_runtime": 0.9921,
-      "eval_samples_per_second": 188.484,
-      "eval_steps_per_second": 3.024,
+      "eval_QUANTITY_precision": 0.4772727272727273,
+      "eval_QUANTITY_recall": 0.7241379310344828,
+      "eval_TIME_f1": 0.9014084507042254,
+      "eval_TIME_number": 34,
+      "eval_TIME_precision": 0.8648648648648649,
+      "eval_TIME_recall": 0.9411764705882353,
+      "eval_loss": 0.17346060276031494,
+      "eval_overall_accuracy": 0.9428993354664041,
+      "eval_overall_f1": 0.7506361323155216,
+      "eval_overall_precision": 0.6781609195402298,
+      "eval_overall_recall": 0.8404558404558404,
+      "eval_runtime": 0.3877,
+      "eval_samples_per_second": 482.308,
+      "eval_steps_per_second": 7.738,
       "step": 424
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1.1405649185180664,
+      "grad_norm": 1.0773650407791138,
       "learning_rate": 4.75e-05,
-      "loss": 0.1869,
+      "loss": 0.1791,
       "step": 530
     },
     {
       "epoch": 5.0,
-      "eval_LOCATION_f1": 0.7682926829268293,
+      "eval_LOCATION_f1": 0.7151515151515151,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6923076923076923,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.6296296296296297,
+      "eval_LOCATION_precision": 0.6413043478260869,
+      "eval_LOCATION_recall": 0.8082191780821918,
+      "eval_ORGANIZATION_f1": 0.6103896103896104,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5257731958762887,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8490566037735849,
+      "eval_ORGANIZATION_precision": 0.5280898876404494,
+      "eval_ORGANIZATION_recall": 0.7230769230769231,
+      "eval_PERSON_f1": 0.8643533123028392,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8035714285714286,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.6086956521739131,
+      "eval_PERSON_precision": 0.8203592814371258,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.591549295774648,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.525,
+      "eval_QUANTITY_precision": 0.5,
       "eval_QUANTITY_recall": 0.7241379310344828,
-      "eval_TIME_f1": 0.7792207792207793,
+      "eval_TIME_f1": 0.7848101265822784,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6976744186046512,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15994760394096375,
-      "eval_overall_accuracy": 0.9473295594388383,
-      "eval_overall_f1": 0.759493670886076,
-      "eval_overall_precision": 0.683371298405467,
-      "eval_overall_recall": 0.8547008547008547,
-      "eval_runtime": 0.9628,
-      "eval_samples_per_second": 194.219,
-      "eval_steps_per_second": 3.116,
+      "eval_TIME_precision": 0.6888888888888889,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.1583884358406067,
+      "eval_overall_accuracy": 0.9456066945606695,
+      "eval_overall_f1": 0.7506361323155216,
+      "eval_overall_precision": 0.6781609195402298,
+      "eval_overall_recall": 0.8404558404558404,
+      "eval_runtime": 0.3854,
+      "eval_samples_per_second": 485.195,
+      "eval_steps_per_second": 7.784,
       "step": 530
     },
     {
       "epoch": 6.0,
-      "grad_norm": 1.9796879291534424,
+      "grad_norm": 1.4482425451278687,
       "learning_rate": 4.7e-05,
-      "loss": 0.169,
+      "loss": 0.1602,
       "step": 636
     },
     {
       "epoch": 6.0,
-      "eval_LOCATION_f1": 0.6993865030674846,
+      "eval_LOCATION_f1": 0.7407407407407406,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6333333333333333,
-      "eval_LOCATION_recall": 0.7808219178082192,
-      "eval_ORGANIZATION_f1": 0.6153846153846154,
+      "eval_LOCATION_precision": 0.6741573033707865,
+      "eval_LOCATION_recall": 0.821917808219178,
+      "eval_ORGANIZATION_f1": 0.6787878787878788,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8769716088328076,
+      "eval_ORGANIZATION_precision": 0.56,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.8633540372670808,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8323353293413174,
+      "eval_PERSON_precision": 0.8081395348837209,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.5897435897435898,
+      "eval_QUANTITY_f1": 0.6153846153846154,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.46938775510204084,
-      "eval_QUANTITY_recall": 0.7931034482758621,
-      "eval_TIME_f1": 0.7073170731707317,
+      "eval_QUANTITY_precision": 0.4897959183673469,
+      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_TIME_f1": 0.7435897435897436,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6041666666666666,
+      "eval_TIME_precision": 0.6590909090909091,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.19615504145622253,
-      "eval_overall_accuracy": 0.9318237755353187,
-      "eval_overall_f1": 0.7416563658838071,
-      "eval_overall_precision": 0.6550218340611353,
-      "eval_overall_recall": 0.8547008547008547,
-      "eval_runtime": 0.9795,
-      "eval_samples_per_second": 190.923,
-      "eval_steps_per_second": 3.063,
+      "eval_loss": 0.18004342913627625,
+      "eval_overall_accuracy": 0.9416687176962836,
+      "eval_overall_f1": 0.7652173913043478,
+      "eval_overall_precision": 0.6784140969162996,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3878,
+      "eval_samples_per_second": 482.177,
+      "eval_steps_per_second": 7.735,
       "step": 636
     },
     {
       "epoch": 7.0,
-      "grad_norm": 0.9608461260795593,
+      "grad_norm": 0.84205162525177,
       "learning_rate": 4.6500000000000005e-05,
-      "loss": 0.154,
+      "loss": 0.1484,
       "step": 742
     },
     {
       "epoch": 7.0,
-      "eval_LOCATION_f1": 0.7560975609756097,
+      "eval_LOCATION_f1": 0.779874213836478,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6813186813186813,
+      "eval_LOCATION_precision": 0.7209302325581395,
       "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.6620689655172414,
+      "eval_ORGANIZATION_f1": 0.6887417218543046,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6,
-      "eval_ORGANIZATION_recall": 0.7384615384615385,
-      "eval_PERSON_f1": 0.8817891373801918,
+      "eval_ORGANIZATION_precision": 0.6046511627906976,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.879746835443038,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8466257668711656,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.6956521739130435,
+      "eval_PERSON_precision": 0.8373493975903614,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.6857142857142856,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6,
+      "eval_QUANTITY_precision": 0.5853658536585366,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7532467532467532,
+      "eval_TIME_f1": 0.8108108108108107,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6744186046511628,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14681145548820496,
-      "eval_overall_accuracy": 0.9492985478710313,
-      "eval_overall_f1": 0.7838541666666667,
-      "eval_overall_precision": 0.7218225419664268,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 0.9965,
-      "eval_samples_per_second": 187.649,
-      "eval_steps_per_second": 3.01,
+      "eval_TIME_precision": 0.75,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.14192061126232147,
+      "eval_overall_accuracy": 0.9549593896135861,
+      "eval_overall_f1": 0.7974025974025974,
+      "eval_overall_precision": 0.7326968973747017,
+      "eval_overall_recall": 0.8746438746438746,
+      "eval_runtime": 0.3862,
+      "eval_samples_per_second": 484.261,
+      "eval_steps_per_second": 7.769,
       "step": 742
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.8899044394493103,
+      "grad_norm": 1.116653561592102,
       "learning_rate": 4.600000000000001e-05,
-      "loss": 0.1439,
+      "loss": 0.1372,
       "step": 848
     },
     {
       "epoch": 8.0,
-      "eval_LOCATION_f1": 0.7354838709677419,
+      "eval_LOCATION_f1": 0.7439024390243901,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6951219512195121,
-      "eval_LOCATION_recall": 0.7808219178082192,
+      "eval_LOCATION_precision": 0.6703296703296703,
+      "eval_LOCATION_recall": 0.8356164383561644,
       "eval_ORGANIZATION_f1": 0.6887417218543046,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.6046511627906976,
       "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.879746835443038,
+      "eval_PERSON_f1": 0.8853503184713376,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8373493975903614,
+      "eval_PERSON_precision": 0.8475609756097561,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.716417910447761,
+      "eval_QUANTITY_f1": 0.6857142857142856,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.631578947368421,
+      "eval_QUANTITY_precision": 0.5853658536585366,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7733333333333334,
+      "eval_TIME_f1": 0.8378378378378379,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7073170731707317,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1457252949476242,
-      "eval_overall_accuracy": 0.9505291656411519,
-      "eval_overall_f1": 0.787958115183246,
-      "eval_overall_precision": 0.7288135593220338,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 1.0009,
-      "eval_samples_per_second": 186.838,
-      "eval_steps_per_second": 2.997,
+      "eval_TIME_precision": 0.775,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.14389163255691528,
+      "eval_overall_accuracy": 0.9527442776273689,
+      "eval_overall_f1": 0.7943078913324709,
+      "eval_overall_precision": 0.7274881516587678,
+      "eval_overall_recall": 0.8746438746438746,
+      "eval_runtime": 0.3852,
+      "eval_samples_per_second": 485.507,
+      "eval_steps_per_second": 7.789,
       "step": 848
     },
     {
       "epoch": 9.0,
-      "grad_norm": 1.0214319229125977,
+      "grad_norm": 1.8104156255722046,
       "learning_rate": 4.55e-05,
-      "loss": 0.1317,
+      "loss": 0.1271,
       "step": 954
     },
     {
       "epoch": 9.0,
-      "eval_LOCATION_f1": 0.7870967741935484,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.6585365853658536,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.6838709677419356,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5454545454545454,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.887459807073955,
+      "eval_ORGANIZATION_precision": 0.5888888888888889,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8571428571428571,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7272727272727273,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7692307692307692,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6486486486486487,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7837837837837837,
-      "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.725,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1383334845304489,
-      "eval_overall_accuracy": 0.9524981540733448,
-      "eval_overall_f1": 0.7948051948051948,
-      "eval_overall_precision": 0.7303102625298329,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 1.0066,
-      "eval_samples_per_second": 185.771,
-      "eval_steps_per_second": 2.98,
+      "eval_QUANTITY_precision": 0.6944444444444444,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8767123287671232,
+      "eval_TIME_number": 34,
+      "eval_TIME_precision": 0.8205128205128205,
+      "eval_TIME_recall": 0.9411764705882353,
+      "eval_loss": 0.12979353964328766,
+      "eval_overall_accuracy": 0.9584051193699237,
+      "eval_overall_f1": 0.8269484808454426,
+      "eval_overall_precision": 0.770935960591133,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3849,
+      "eval_samples_per_second": 485.83,
+      "eval_steps_per_second": 7.794,
       "step": 954
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.8760809302330017,
+      "grad_norm": 0.966005802154541,
       "learning_rate": 4.5e-05,
-      "loss": 0.1261,
+      "loss": 0.1255,
       "step": 1060
     },
     {
@@ -372,154 +372,154 @@
       "eval_LOCATION_number": 73,
       "eval_LOCATION_precision": 0.7875,
       "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.6887417218543046,
+      "eval_ORGANIZATION_f1": 0.6986301369863014,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6046511627906976,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.6296296296296297,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7076923076923076,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7352941176470588,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6388888888888888,
-      "eval_QUANTITY_recall": 0.7931034482758621,
-      "eval_TIME_f1": 0.8219178082191781,
+      "eval_QUANTITY_precision": 0.6410256410256411,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7692307692307693,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1302676945924759,
-      "eval_overall_accuracy": 0.9559438838296825,
-      "eval_overall_f1": 0.8100929614873837,
-      "eval_overall_precision": 0.7587064676616916,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9994,
-      "eval_samples_per_second": 187.11,
-      "eval_steps_per_second": 3.002,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.12757907807826996,
+      "eval_overall_accuracy": 0.9601279842480925,
+      "eval_overall_f1": 0.8284182305630027,
+      "eval_overall_precision": 0.7822784810126582,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3839,
+      "eval_samples_per_second": 487.093,
+      "eval_steps_per_second": 7.814,
       "step": 1060
     },
     {
       "epoch": 11.0,
-      "grad_norm": 0.8754971027374268,
+      "grad_norm": 0.9226321578025818,
       "learning_rate": 4.4500000000000004e-05,
-      "loss": 0.1191,
+      "loss": 0.118,
       "step": 1166
     },
     {
       "epoch": 11.0,
-      "eval_LOCATION_f1": 0.781456953642384,
+      "eval_LOCATION_f1": 0.8051948051948052,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7564102564102564,
-      "eval_LOCATION_recall": 0.8082191780821918,
-      "eval_ORGANIZATION_f1": 0.7234042553191491,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.697986577181208,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6710526315789473,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.887459807073955,
+      "eval_ORGANIZATION_precision": 0.6190476190476191,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8867313915857605,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8571428571428571,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.676923076923077,
+      "eval_PERSON_precision": 0.8616352201257862,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7352941176470588,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6111111111111112,
-      "eval_QUANTITY_recall": 0.7586206896551724,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.6410256410256411,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.12534178793430328,
-      "eval_overall_accuracy": 0.9596357371400444,
-      "eval_overall_f1": 0.8070175438596491,
-      "eval_overall_precision": 0.7666666666666667,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9868,
-      "eval_samples_per_second": 189.5,
-      "eval_steps_per_second": 3.04,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.13041509687900543,
+      "eval_overall_accuracy": 0.9584051193699237,
+      "eval_overall_f1": 0.8138297872340426,
+      "eval_overall_precision": 0.7630922693266833,
+      "eval_overall_recall": 0.8717948717948718,
+      "eval_runtime": 0.3859,
+      "eval_samples_per_second": 484.593,
+      "eval_steps_per_second": 7.774,
       "step": 1166
     },
     {
       "epoch": 12.0,
-      "grad_norm": 0.8746627569198608,
+      "grad_norm": 0.9765021800994873,
       "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.1131,
+      "loss": 0.1125,
       "step": 1272
     },
     {
       "epoch": 12.0,
-      "eval_LOCATION_f1": 0.7361963190184049,
+      "eval_LOCATION_f1": 0.7848101265822783,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.6666666666666666,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7051282051282052,
+      "eval_LOCATION_precision": 0.7294117647058823,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.738255033557047,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6043956043956044,
+      "eval_ORGANIZATION_precision": 0.6547619047619048,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8910256410256411,
+      "eval_PERSON_f1": 0.8860759493670886,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8580246913580247,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7575757575757576,
+      "eval_PERSON_precision": 0.8433734939759037,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7352941176470588,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6756756756756757,
+      "eval_QUANTITY_precision": 0.6410256410256411,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.7532467532467532,
+      "eval_TIME_f1": 0.8,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6744186046511628,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1429539918899536,
-      "eval_overall_accuracy": 0.9537287718434654,
-      "eval_overall_f1": 0.7958656330749354,
-      "eval_overall_precision": 0.7281323877068558,
-      "eval_overall_recall": 0.8774928774928775,
-      "eval_runtime": 1.0275,
-      "eval_samples_per_second": 181.996,
-      "eval_steps_per_second": 2.92,
+      "eval_TIME_precision": 0.7317073170731707,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.14262500405311584,
+      "eval_overall_accuracy": 0.9559438838296825,
+      "eval_overall_f1": 0.8146214099216709,
+      "eval_overall_precision": 0.7518072289156627,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3854,
+      "eval_samples_per_second": 485.157,
+      "eval_steps_per_second": 7.783,
       "step": 1272
     },
     {
       "epoch": 13.0,
-      "grad_norm": 0.7008482813835144,
+      "grad_norm": 0.696005642414093,
       "learning_rate": 4.35e-05,
-      "loss": 0.1101,
+      "loss": 0.1032,
       "step": 1378
     },
     {
       "epoch": 13.0,
-      "eval_LOCATION_f1": 0.7999999999999999,
+      "eval_LOCATION_f1": 0.8079470198675496,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7792207792207793,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.728476821192053,
+      "eval_LOCATION_precision": 0.782051282051282,
+      "eval_LOCATION_recall": 0.8356164383561644,
+      "eval_ORGANIZATION_f1": 0.713375796178344,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6395348837209303,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8938906752411575,
+      "eval_ORGANIZATION_precision": 0.6086956521739131,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8633540372670807,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.6376811594202898,
+      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.746268656716418,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.55,
-      "eval_QUANTITY_recall": 0.7586206896551724,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.6578947368421053,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8823529411764706,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.13461139798164368,
-      "eval_overall_accuracy": 0.9564361309377307,
-      "eval_overall_f1": 0.8090185676392573,
-      "eval_overall_precision": 0.7568238213399504,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9921,
-      "eval_samples_per_second": 188.484,
-      "eval_steps_per_second": 3.024,
+      "eval_TIME_precision": 0.8823529411764706,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.13183027505874634,
+      "eval_overall_accuracy": 0.9593896135860202,
+      "eval_overall_f1": 0.8255659121171771,
+      "eval_overall_precision": 0.775,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3861,
+      "eval_samples_per_second": 484.32,
+      "eval_steps_per_second": 7.77,
       "step": 1378
     },
     {
       "epoch": 14.0,
-      "grad_norm": 0.6436917185783386,
+      "grad_norm": 0.6964078545570374,
       "learning_rate": 4.3e-05,
-      "loss": 0.105,
+      "loss": 0.1046,
       "step": 1484
     },
     {
@@ -528,739 +528,739 @@
       "eval_LOCATION_number": 73,
       "eval_LOCATION_precision": 0.7294117647058823,
       "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7234042553191491,
+      "eval_ORGANIZATION_f1": 0.7105263157894737,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6710526315789473,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8932038834951457,
+      "eval_ORGANIZATION_precision": 0.6206896551724138,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8679245283018868,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.746268656716418,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7536231884057972,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6578947368421053,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.7837837837837837,
+      "eval_QUANTITY_precision": 0.65,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.7368421052631577,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.725,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.13453252613544464,
-      "eval_overall_accuracy": 0.9591434900319961,
-      "eval_overall_f1": 0.814419225634179,
-      "eval_overall_precision": 0.7663316582914573,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9965,
-      "eval_samples_per_second": 187.656,
-      "eval_steps_per_second": 3.011,
+      "eval_TIME_precision": 0.6666666666666666,
+      "eval_TIME_recall": 0.8235294117647058,
+      "eval_loss": 0.1450473964214325,
+      "eval_overall_accuracy": 0.9564361309377307,
+      "eval_overall_f1": 0.807843137254902,
+      "eval_overall_precision": 0.7463768115942029,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3849,
+      "eval_samples_per_second": 485.803,
+      "eval_steps_per_second": 7.794,
       "step": 1484
     },
     {
       "epoch": 15.0,
-      "grad_norm": 1.208773136138916,
+      "grad_norm": 1.2660812139511108,
       "learning_rate": 4.25e-05,
-      "loss": 0.1043,
+      "loss": 0.0969,
       "step": 1590
     },
     {
       "epoch": 15.0,
-      "eval_LOCATION_f1": 0.7843137254901962,
+      "eval_LOCATION_f1": 0.8129032258064516,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.75,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7682926829268293,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7162162162162162,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
+      "eval_ORGANIZATION_precision": 0.6385542168674698,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7575757575757576,
+      "eval_PERSON_precision": 0.8726114649681529,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7692307692307692,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6756756756756757,
+      "eval_QUANTITY_precision": 0.6944444444444444,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1265690177679062,
-      "eval_overall_accuracy": 0.961112478464189,
-      "eval_overall_f1": 0.8221024258760108,
-      "eval_overall_precision": 0.7800511508951407,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 1.0073,
-      "eval_samples_per_second": 185.651,
-      "eval_steps_per_second": 2.978,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.13275112211704254,
+      "eval_overall_accuracy": 0.9618508491262614,
+      "eval_overall_f1": 0.8284182305630027,
+      "eval_overall_precision": 0.7822784810126582,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3921,
+      "eval_samples_per_second": 476.967,
+      "eval_steps_per_second": 7.652,
       "step": 1590
     },
     {
       "epoch": 16.0,
-      "grad_norm": 1.1977620124816895,
+      "grad_norm": 1.0079460144042969,
       "learning_rate": 4.2e-05,
-      "loss": 0.0979,
+      "loss": 0.0916,
       "step": 1696
     },
     {
       "epoch": 16.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7236842105263158,
+      "eval_LOCATION_precision": 0.8,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7323943661971831,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.632183908045977,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.90032154340836,
+      "eval_ORGANIZATION_precision": 0.6753246753246753,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8974358974358974,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_precision": 0.8641975308641975,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.6666666666666666,
+      "eval_QUANTITY_f1": 0.7575757575757576,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.5581395348837209,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7368421052631577,
+      "eval_QUANTITY_precision": 0.6756756756756757,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8169014084507041,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6666666666666666,
-      "eval_TIME_recall": 0.8235294117647058,
-      "eval_loss": 0.1487545669078827,
-      "eval_overall_accuracy": 0.9527442776273689,
-      "eval_overall_f1": 0.8026143790849672,
-      "eval_overall_precision": 0.7415458937198067,
-      "eval_overall_recall": 0.8746438746438746,
-      "eval_runtime": 0.9958,
-      "eval_samples_per_second": 187.787,
-      "eval_steps_per_second": 3.013,
+      "eval_TIME_precision": 0.7837837837837838,
+      "eval_TIME_recall": 0.8529411764705882,
+      "eval_loss": 0.1358853280544281,
+      "eval_overall_accuracy": 0.9625892197883338,
+      "eval_overall_f1": 0.8333333333333334,
+      "eval_overall_precision": 0.7888040712468194,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3926,
+      "eval_samples_per_second": 476.356,
+      "eval_steps_per_second": 7.642,
       "step": 1696
     },
     {
       "epoch": 17.0,
-      "grad_norm": 0.41393932700157166,
+      "grad_norm": 2.586317539215088,
       "learning_rate": 4.15e-05,
-      "loss": 0.0954,
+      "loss": 0.0915,
       "step": 1802
     },
     {
       "epoch": 17.0,
-      "eval_LOCATION_f1": 0.8104575163398693,
+      "eval_LOCATION_f1": 0.8076923076923077,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.775,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7027027027027027,
+      "eval_LOCATION_precision": 0.7590361445783133,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7310344827586207,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6265060240963856,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.90032154340836,
+      "eval_ORGANIZATION_precision": 0.6625,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_precision": 0.875,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7631578947368423,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.6904761904761905,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.13488559424877167,
-      "eval_overall_accuracy": 0.958897366477972,
-      "eval_overall_f1": 0.8175765645805593,
-      "eval_overall_precision": 0.7675,
-      "eval_overall_recall": 0.8746438746438746,
-      "eval_runtime": 0.9952,
-      "eval_samples_per_second": 187.902,
-      "eval_steps_per_second": 3.014,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.1329522430896759,
+      "eval_overall_accuracy": 0.9623430962343096,
+      "eval_overall_f1": 0.8398384925975775,
+      "eval_overall_precision": 0.7959183673469388,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3867,
+      "eval_samples_per_second": 483.521,
+      "eval_steps_per_second": 7.757,
       "step": 1802
     },
     {
       "epoch": 18.0,
-      "grad_norm": 1.0734295845031738,
+      "grad_norm": 0.9525734782218933,
       "learning_rate": 4.1e-05,
-      "loss": 0.0928,
+      "loss": 0.0917,
       "step": 1908
     },
     {
       "epoch": 18.0,
-      "eval_LOCATION_f1": 0.7870967741935484,
+      "eval_LOCATION_f1": 0.8289473684210525,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7439024390243902,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.726027397260274,
+      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7310344827586207,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.654320987654321,
+      "eval_ORGANIZATION_precision": 0.6625,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7692307692307692,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.716417910447761,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6944444444444444,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.7837837837837837,
+      "eval_QUANTITY_precision": 0.631578947368421,
+      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.725,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1342119574546814,
-      "eval_overall_accuracy": 0.9596357371400444,
-      "eval_overall_f1": 0.8133333333333332,
-      "eval_overall_precision": 0.7644110275689223,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 1.0888,
-      "eval_samples_per_second": 171.747,
-      "eval_steps_per_second": 2.755,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.14254941046237946,
+      "eval_overall_accuracy": 0.9591434900319961,
+      "eval_overall_f1": 0.8279569892473119,
+      "eval_overall_precision": 0.7837150127226463,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3875,
+      "eval_samples_per_second": 482.589,
+      "eval_steps_per_second": 7.742,
       "step": 1908
     },
     {
       "epoch": 19.0,
-      "grad_norm": 0.5558896660804749,
+      "grad_norm": 0.7584782242774963,
       "learning_rate": 4.05e-05,
-      "loss": 0.0884,
+      "loss": 0.0883,
       "step": 2014
     },
     {
       "epoch": 19.0,
-      "eval_LOCATION_f1": 0.8157894736842105,
+      "eval_LOCATION_f1": 0.8289473684210525,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7042253521126761,
+      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7586206896551724,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6493506493506493,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6875,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_precision": 0.8625,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_QUANTITY_f1": 0.8125,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.7428571428571429,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
+      "eval_TIME_precision": 0.7631578947368421,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.12890850007534027,
-      "eval_overall_accuracy": 0.9616047255722373,
-      "eval_overall_f1": 0.8200270635994589,
-      "eval_overall_precision": 0.7809278350515464,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9958,
-      "eval_samples_per_second": 187.787,
-      "eval_steps_per_second": 3.013,
+      "eval_loss": 0.13237319886684418,
+      "eval_overall_accuracy": 0.963081466896382,
+      "eval_overall_f1": 0.8371467025572006,
+      "eval_overall_precision": 0.7933673469387755,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3917,
+      "eval_samples_per_second": 477.449,
+      "eval_steps_per_second": 7.66,
       "step": 2014
     },
     {
       "epoch": 20.0,
-      "grad_norm": 0.4977124333381653,
+      "grad_norm": 0.6288883090019226,
       "learning_rate": 4e-05,
-      "loss": 0.0851,
+      "loss": 0.0838,
       "step": 2120
     },
     {
       "epoch": 20.0,
-      "eval_LOCATION_f1": 0.7692307692307693,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7228915662650602,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7412587412587412,
+      "eval_LOCATION_precision": 0.7901234567901234,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6794871794871795,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8903225806451613,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8625,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.6857142857142856,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
+      "eval_QUANTITY_precision": 0.5853658536585366,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7837837837837837,
+      "eval_TIME_f1": 0.8985507246376812,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.725,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.137658029794693,
-      "eval_overall_accuracy": 0.9616047255722373,
-      "eval_overall_f1": 0.8150134048257373,
-      "eval_overall_precision": 0.769620253164557,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9955,
-      "eval_samples_per_second": 187.845,
-      "eval_steps_per_second": 3.014,
+      "eval_TIME_precision": 0.8857142857142857,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.1400074064731598,
+      "eval_overall_accuracy": 0.9613586020182131,
+      "eval_overall_f1": 0.8364611260053618,
+      "eval_overall_precision": 0.789873417721519,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3853,
+      "eval_samples_per_second": 485.372,
+      "eval_steps_per_second": 7.787,
       "step": 2120
     },
     {
       "epoch": 21.0,
-      "grad_norm": 2.0087778568267822,
+      "grad_norm": 2.4087135791778564,
       "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.0819,
+      "loss": 0.0809,
       "step": 2226
     },
     {
       "epoch": 21.0,
-      "eval_LOCATION_f1": 0.8235294117647058,
+      "eval_LOCATION_f1": 0.8289473684210525,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7875,
+      "eval_LOCATION_precision": 0.7974683544303798,
       "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_ORGANIZATION_f1": 0.7466666666666668,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_ORGANIZATION_precision": 0.6588235294117647,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.8846153846153846,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_precision": 0.8518518518518519,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_QUANTITY_f1": 0.7419354838709677,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.696969696969697,
+      "eval_QUANTITY_recall": 0.7931034482758621,
+      "eval_TIME_f1": 0.8985507246376812,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.12961311638355255,
-      "eval_overall_accuracy": 0.9633275904504061,
-      "eval_overall_f1": 0.8319783197831978,
-      "eval_overall_precision": 0.7932816537467701,
-      "eval_overall_recall": 0.8746438746438746,
-      "eval_runtime": 1.0079,
-      "eval_samples_per_second": 185.535,
-      "eval_steps_per_second": 2.976,
+      "eval_TIME_precision": 0.8857142857142857,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.13903410732746124,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8348993288590604,
+      "eval_overall_precision": 0.7893401015228426,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3889,
+      "eval_samples_per_second": 480.805,
+      "eval_steps_per_second": 7.713,
       "step": 2226
     },
     {
       "epoch": 22.0,
-      "grad_norm": 1.2587146759033203,
+      "grad_norm": 1.5907405614852905,
       "learning_rate": 3.9000000000000006e-05,
-      "loss": 0.0858,
+      "loss": 0.0795,
       "step": 2332
     },
     {
       "epoch": 22.0,
-      "eval_LOCATION_f1": 0.7999999999999999,
+      "eval_LOCATION_f1": 0.8421052631578948,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7792207792207793,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7199999999999999,
+      "eval_LOCATION_precision": 0.810126582278481,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7448275862068966,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6352941176470588,
+      "eval_ORGANIZATION_precision": 0.675,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.8064516129032258,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.75,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7575757575757576,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.6857142857142857,
+      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_TIME_f1": 0.8169014084507041,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
+      "eval_TIME_precision": 0.7837837837837838,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.13012367486953735,
+      "eval_loss": 0.13828279078006744,
       "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8281461434370772,
-      "eval_overall_precision": 0.788659793814433,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9889,
-      "eval_samples_per_second": 189.099,
-      "eval_steps_per_second": 3.034,
+      "eval_overall_f1": 0.8378378378378378,
+      "eval_overall_precision": 0.7969151670951157,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3894,
+      "eval_samples_per_second": 480.224,
+      "eval_steps_per_second": 7.704,
       "step": 2332
     },
     {
       "epoch": 23.0,
-      "grad_norm": 1.0316872596740723,
+      "grad_norm": 1.1399990320205688,
       "learning_rate": 3.85e-05,
-      "loss": 0.0789,
+      "loss": 0.0778,
       "step": 2438
     },
     {
       "epoch": 23.0,
-      "eval_LOCATION_f1": 0.7922077922077922,
+      "eval_LOCATION_f1": 0.8441558441558442,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7530864197530864,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7391304347826088,
+      "eval_LOCATION_precision": 0.8024691358024691,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7448275862068966,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6986301369863014,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.9061488673139159,
+      "eval_ORGANIZATION_precision": 0.675,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_precision": 0.875,
       "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.7936507936507937,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7352941176470589,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8169014084507041,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7837837837837838,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.13557566702365875,
-      "eval_overall_accuracy": 0.9628353433423579,
-      "eval_overall_f1": 0.8326530612244898,
-      "eval_overall_precision": 0.796875,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9917,
-      "eval_samples_per_second": 188.567,
-      "eval_steps_per_second": 3.025,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.1377546340227127,
+      "eval_overall_accuracy": 0.9645582082205267,
+      "eval_overall_f1": 0.8502024291497975,
+      "eval_overall_precision": 0.8076923076923077,
+      "eval_overall_recall": 0.8974358974358975,
+      "eval_runtime": 0.3909,
+      "eval_samples_per_second": 478.381,
+      "eval_steps_per_second": 7.675,
       "step": 2438
     },
     {
       "epoch": 24.0,
-      "grad_norm": 0.8608385324478149,
+      "grad_norm": 0.7097505331039429,
       "learning_rate": 3.8e-05,
-      "loss": 0.0763,
+      "loss": 0.0731,
       "step": 2544
     },
     {
       "epoch": 24.0,
-      "eval_LOCATION_f1": 0.7894736842105263,
+      "eval_LOCATION_f1": 0.8311688311688312,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.759493670886076,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7162162162162162,
+      "eval_LOCATION_precision": 0.7901234567901234,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7417218543046358,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6385542168674698,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.9025974025974027,
+      "eval_ORGANIZATION_precision": 0.6511627906976745,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.90032154340836,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.879746835443038,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.8064516129032258,
+      "eval_PERSON_precision": 0.8695652173913043,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7575757575757576,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8695652173913043,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1364046037197113,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8236877523553163,
-      "eval_overall_precision": 0.7806122448979592,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9754,
-      "eval_samples_per_second": 191.709,
-      "eval_steps_per_second": 3.076,
+      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.14163164794445038,
+      "eval_overall_accuracy": 0.9625892197883338,
+      "eval_overall_f1": 0.8449197860962567,
+      "eval_overall_precision": 0.7959697732997482,
+      "eval_overall_recall": 0.9002849002849003,
+      "eval_runtime": 0.3873,
+      "eval_samples_per_second": 482.826,
+      "eval_steps_per_second": 7.746,
       "step": 2544
     },
     {
       "epoch": 25.0,
-      "grad_norm": 0.9957919716835022,
+      "grad_norm": 0.6983250379562378,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0734,
+      "loss": 0.0729,
       "step": 2650
     },
     {
       "epoch": 25.0,
-      "eval_LOCATION_f1": 0.7712418300653595,
+      "eval_LOCATION_f1": 0.8289473684210525,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7375,
-      "eval_LOCATION_recall": 0.8082191780821918,
-      "eval_ORGANIZATION_f1": 0.7375886524822695,
+      "eval_LOCATION_precision": 0.7974683544303798,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7586206896551724,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6842105263157895,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.9032258064516129,
+      "eval_ORGANIZATION_precision": 0.6875,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8910256410256411,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.875,
-      "eval_PERSON_recall": 0.9333333333333333,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8580246913580247,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.6956521739130435,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
+      "eval_QUANTITY_precision": 0.6,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7733333333333334,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7073170731707317,
+      "eval_TIME_precision": 0.7631578947368421,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1437826156616211,
-      "eval_overall_accuracy": 0.961112478464189,
-      "eval_overall_f1": 0.8183041722745626,
-      "eval_overall_precision": 0.7755102040816326,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9957,
-      "eval_samples_per_second": 187.814,
-      "eval_steps_per_second": 3.013,
+      "eval_loss": 0.15113840997219086,
+      "eval_overall_accuracy": 0.9608663549101649,
+      "eval_overall_f1": 0.8266666666666667,
+      "eval_overall_precision": 0.7769423558897243,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3857,
+      "eval_samples_per_second": 484.853,
+      "eval_steps_per_second": 7.778,
       "step": 2650
     },
     {
       "epoch": 26.0,
-      "grad_norm": 0.6454270482063293,
+      "grad_norm": 0.6798135638237,
       "learning_rate": 3.7e-05,
-      "loss": 0.0731,
+      "loss": 0.0724,
       "step": 2756
     },
     {
       "epoch": 26.0,
-      "eval_LOCATION_f1": 0.7712418300653595,
+      "eval_LOCATION_f1": 0.8211920529801324,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7375,
-      "eval_LOCATION_recall": 0.8082191780821918,
-      "eval_ORGANIZATION_f1": 0.7092198581560285,
+      "eval_LOCATION_precision": 0.7948717948717948,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.7714285714285715,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6578947368421053,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.896774193548387,
+      "eval_ORGANIZATION_precision": 0.72,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.86875,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7272727272727273,
+      "eval_PERSON_precision": 0.8670886075949367,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.676056338028169,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6486486486486487,
+      "eval_QUANTITY_precision": 0.5714285714285714,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_TIME_f1": 0.7837837837837837,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
+      "eval_TIME_precision": 0.725,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14986544847488403,
+      "eval_loss": 0.14874529838562012,
       "eval_overall_accuracy": 0.958897366477972,
-      "eval_overall_f1": 0.8102288021534321,
-      "eval_overall_precision": 0.7678571428571429,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 0.9887,
-      "eval_samples_per_second": 189.131,
-      "eval_steps_per_second": 3.034,
+      "eval_overall_f1": 0.8225806451612903,
+      "eval_overall_precision": 0.7786259541984732,
+      "eval_overall_recall": 0.8717948717948718,
+      "eval_runtime": 0.386,
+      "eval_samples_per_second": 484.512,
+      "eval_steps_per_second": 7.773,
       "step": 2756
     },
     {
       "epoch": 27.0,
-      "grad_norm": 0.7397086024284363,
+      "grad_norm": 1.0639210939407349,
       "learning_rate": 3.65e-05,
-      "loss": 0.0704,
+      "loss": 0.0699,
       "step": 2862
     },
     {
       "epoch": 27.0,
-      "eval_LOCATION_f1": 0.7973856209150326,
+      "eval_LOCATION_f1": 0.802547770700637,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7625,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7375886524822695,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7724137931034483,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6842105263157895,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8853503184713376,
+      "eval_ORGANIZATION_precision": 0.7,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8475609756097561,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.716417910447761,
+      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.6666666666666666,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.631578947368421,
+      "eval_QUANTITY_precision": 0.5581395348837209,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.767123287671233,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.717948717948718,
-      "eval_TIME_recall": 0.8235294117647058,
-      "eval_loss": 0.1502029001712799,
-      "eval_overall_accuracy": 0.9591434900319961,
-      "eval_overall_f1": 0.8128342245989305,
-      "eval_overall_precision": 0.7657430730478589,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9947,
-      "eval_samples_per_second": 188.003,
-      "eval_steps_per_second": 3.016,
+      "eval_TIME_precision": 0.7631578947368421,
+      "eval_TIME_recall": 0.8529411764705882,
+      "eval_loss": 0.14704443514347076,
+      "eval_overall_accuracy": 0.9601279842480925,
+      "eval_overall_f1": 0.8253968253968255,
+      "eval_overall_precision": 0.7703703703703704,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3902,
+      "eval_samples_per_second": 479.247,
+      "eval_steps_per_second": 7.688,
       "step": 2862
     },
     {
       "epoch": 28.0,
-      "grad_norm": 1.38124418258667,
+      "grad_norm": 1.30838942527771,
       "learning_rate": 3.6e-05,
-      "loss": 0.0679,
+      "loss": 0.0669,
       "step": 2968
     },
     {
       "epoch": 28.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.8243243243243243,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.75,
+      "eval_LOCATION_precision": 0.8133333333333334,
+      "eval_LOCATION_recall": 0.8356164383561644,
+      "eval_ORGANIZATION_f1": 0.702290076335878,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6835443037974683,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_ORGANIZATION_precision": 0.696969696969697,
+      "eval_ORGANIZATION_recall": 0.7076923076923077,
       "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.8742138364779874,
       "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.71875,
+      "eval_QUANTITY_f1": 0.75,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6571428571428571,
-      "eval_QUANTITY_recall": 0.7931034482758621,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.6857142857142857,
+      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14466358721256256,
-      "eval_overall_accuracy": 0.9618508491262614,
-      "eval_overall_f1": 0.8232118758434548,
-      "eval_overall_precision": 0.782051282051282,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9878,
-      "eval_samples_per_second": 189.3,
-      "eval_steps_per_second": 3.037,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.136509507894516,
+      "eval_overall_accuracy": 0.9633275904504061,
+      "eval_overall_f1": 0.8287292817679558,
+      "eval_overall_precision": 0.8042895442359249,
+      "eval_overall_recall": 0.8547008547008547,
+      "eval_runtime": 0.3858,
+      "eval_samples_per_second": 484.693,
+      "eval_steps_per_second": 7.776,
       "step": 2968
     },
     {
       "epoch": 29.0,
-      "grad_norm": 0.8041614294052124,
+      "grad_norm": 3.5960898399353027,
       "learning_rate": 3.55e-05,
-      "loss": 0.0657,
+      "loss": 0.0671,
       "step": 3074
     },
     {
       "epoch": 29.0,
-      "eval_LOCATION_f1": 0.8211920529801324,
+      "eval_LOCATION_f1": 0.847682119205298,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7948717948717948,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7346938775510204,
+      "eval_LOCATION_precision": 0.8205128205128205,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7746478873239436,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6585365853658537,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.894736842105263,
+      "eval_ORGANIZATION_precision": 0.7142857142857143,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8938906752411575,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8831168831168831,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.8633540372670807,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.9117647058823528,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8695652173913043,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.9117647058823529,
-      "eval_TIME_recall": 0.9117647058823529,
-      "eval_loss": 0.1336374580860138,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.8376534788540245,
-      "eval_overall_precision": 0.8036649214659686,
-      "eval_overall_recall": 0.8746438746438746,
-      "eval_runtime": 0.9977,
-      "eval_samples_per_second": 187.429,
-      "eval_steps_per_second": 3.007,
+      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.13075676560401917,
+      "eval_overall_accuracy": 0.9667733202067438,
+      "eval_overall_f1": 0.8532608695652173,
+      "eval_overall_precision": 0.8155844155844156,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3861,
+      "eval_samples_per_second": 484.295,
+      "eval_steps_per_second": 7.769,
       "step": 3074
     },
     {
       "epoch": 30.0,
-      "grad_norm": 0.9247944951057434,
+      "grad_norm": 1.5083116292953491,
       "learning_rate": 3.5e-05,
-      "loss": 0.067,
+      "loss": 0.0645,
       "step": 3180
     },
     {
       "epoch": 30.0,
-      "eval_LOCATION_f1": 0.7919463087248322,
+      "eval_LOCATION_f1": 0.8187919463087249,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7763157894736842,
-      "eval_LOCATION_recall": 0.8082191780821918,
-      "eval_ORGANIZATION_f1": 0.7210884353741497,
+      "eval_LOCATION_precision": 0.8026315789473685,
+      "eval_LOCATION_recall": 0.8356164383561644,
+      "eval_ORGANIZATION_f1": 0.75,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6463414634146342,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8990228013029316,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8789808917197452,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8169014084507041,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
+      "eval_TIME_precision": 0.7837837837837838,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14372512698173523,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8222523744911805,
-      "eval_overall_precision": 0.7849740932642487,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 1.0038,
-      "eval_samples_per_second": 186.284,
-      "eval_steps_per_second": 2.989,
+      "eval_loss": 0.1309884488582611,
+      "eval_overall_accuracy": 0.9657888259906473,
+      "eval_overall_f1": 0.8423913043478262,
+      "eval_overall_precision": 0.8051948051948052,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3879,
+      "eval_samples_per_second": 482.094,
+      "eval_steps_per_second": 7.734,
       "step": 3180
     },
     {
       "epoch": 31.0,
-      "grad_norm": 0.41009989380836487,
+      "grad_norm": 0.48864850401878357,
       "learning_rate": 3.45e-05,
-      "loss": 0.0649,
+      "loss": 0.0598,
       "step": 3286
     },
     {
       "epoch": 31.0,
-      "eval_LOCATION_f1": 0.8205128205128205,
+      "eval_LOCATION_f1": 0.8101265822784809,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7710843373493976,
+      "eval_LOCATION_precision": 0.7529411764705882,
       "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_ORGANIZATION_f1": 0.7517730496453899,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.9019607843137256,
+      "eval_ORGANIZATION_precision": 0.6973684210526315,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8846153846153846,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7777777777777778,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8285714285714286,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7368421052631579,
-      "eval_TIME_recall": 0.8235294117647058,
-      "eval_loss": 0.1421852558851242,
-      "eval_overall_accuracy": 0.9625892197883338,
-      "eval_overall_f1": 0.8283378746594006,
-      "eval_overall_precision": 0.793733681462141,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9943,
-      "eval_samples_per_second": 188.063,
-      "eval_steps_per_second": 3.017,
+      "eval_TIME_precision": 0.8055555555555556,
+      "eval_TIME_recall": 0.8529411764705882,
+      "eval_loss": 0.13982480764389038,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8367071524966261,
+      "eval_overall_precision": 0.7948717948717948,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3914,
+      "eval_samples_per_second": 477.78,
+      "eval_steps_per_second": 7.665,
       "step": 3286
     },
     {
       "epoch": 32.0,
-      "grad_norm": 0.4342913031578064,
+      "grad_norm": 0.22911441326141357,
       "learning_rate": 3.4000000000000007e-05,
-      "loss": 0.0619,
+      "loss": 0.0598,
       "step": 3392
     },
     {
       "epoch": 32.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.802547770700637,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7236842105263158,
+      "eval_LOCATION_precision": 0.75,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7534246575342466,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.632183908045977,
+      "eval_ORGANIZATION_precision": 0.6790123456790124,
       "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.896103896103896,
+      "eval_PERSON_f1": 0.8938906752411575,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8734177215189873,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8633540372670807,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7272727272727273,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
+      "eval_QUANTITY_precision": 0.6486486486486487,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1523984968662262,
-      "eval_overall_accuracy": 0.9593896135860202,
-      "eval_overall_f1": 0.8208556149732621,
-      "eval_overall_precision": 0.7732997481108312,
-      "eval_overall_recall": 0.8746438746438746,
-      "eval_runtime": 1.0018,
-      "eval_samples_per_second": 186.663,
-      "eval_steps_per_second": 2.995,
+      "eval_TIME_precision": 0.7631578947368421,
+      "eval_TIME_recall": 0.8529411764705882,
+      "eval_loss": 0.14981552958488464,
+      "eval_overall_accuracy": 0.9608663549101649,
+      "eval_overall_f1": 0.824468085106383,
+      "eval_overall_precision": 0.773067331670823,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3886,
+      "eval_samples_per_second": 481.274,
+      "eval_steps_per_second": 7.721,
       "step": 3392
     },
     {
       "epoch": 33.0,
-      "grad_norm": 0.8614174127578735,
+      "grad_norm": 0.9966171979904175,
       "learning_rate": 3.35e-05,
-      "loss": 0.0616,
+      "loss": 0.0607,
       "step": 3498
     },
     {
@@ -1269,2653 +1269,2653 @@
       "eval_LOCATION_number": 73,
       "eval_LOCATION_precision": 0.7922077922077922,
       "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.72992700729927,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6944444444444444,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8910256410256411,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8580246913580247,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
       "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
       "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.14742247760295868,
-      "eval_overall_accuracy": 0.9623430962343096,
-      "eval_overall_f1": 0.8273972602739725,
-      "eval_overall_precision": 0.7968337730870713,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9963,
-      "eval_samples_per_second": 187.698,
-      "eval_steps_per_second": 3.011,
+      "eval_loss": 0.14727307856082916,
+      "eval_overall_accuracy": 0.9625892197883338,
+      "eval_overall_f1": 0.8389715832205684,
+      "eval_overall_precision": 0.7989690721649485,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3876,
+      "eval_samples_per_second": 482.477,
+      "eval_steps_per_second": 7.74,
       "step": 3498
     },
     {
       "epoch": 34.0,
-      "grad_norm": 1.166977882385254,
+      "grad_norm": 1.1382331848144531,
       "learning_rate": 3.3e-05,
-      "loss": 0.0597,
+      "loss": 0.0583,
       "step": 3604
     },
     {
       "epoch": 34.0,
-      "eval_LOCATION_f1": 0.8181818181818182,
+      "eval_LOCATION_f1": 0.8400000000000001,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7777777777777778,
+      "eval_LOCATION_precision": 0.8181818181818182,
       "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7333333333333334,
+      "eval_ORGANIZATION_f1": 0.7659574468085107,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6470588235294118,
-      "eval_ORGANIZATION_recall": 0.8461538461538461,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_ORGANIZATION_precision": 0.7105263157894737,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7936507936507937,
+      "eval_PERSON_precision": 0.8571428571428571,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7352941176470589,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8055555555555555,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.861111111111111,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7631578947368421,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14954759180545807,
-      "eval_overall_accuracy": 0.958897366477972,
-      "eval_overall_f1": 0.8251001335113485,
-      "eval_overall_precision": 0.7763819095477387,
-      "eval_overall_recall": 0.8803418803418803,
-      "eval_runtime": 1.0032,
-      "eval_samples_per_second": 186.407,
-      "eval_steps_per_second": 2.99,
+      "eval_TIME_precision": 0.8157894736842105,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.14073994755744934,
+      "eval_overall_accuracy": 0.9670194437607679,
+      "eval_overall_f1": 0.8466757123473541,
+      "eval_overall_precision": 0.8082901554404145,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3867,
+      "eval_samples_per_second": 483.635,
+      "eval_steps_per_second": 7.759,
       "step": 3604
     },
     {
       "epoch": 35.0,
-      "grad_norm": 0.7428762912750244,
+      "grad_norm": 0.9959651231765747,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.0594,
+      "loss": 0.0558,
       "step": 3710
     },
     {
       "epoch": 35.0,
-      "eval_LOCATION_f1": 0.7820512820512822,
+      "eval_LOCATION_f1": 0.8051948051948052,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7349397590361446,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.757142857142857,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
+      "eval_ORGANIZATION_precision": 0.7066666666666667,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8753993610223643,
+      "eval_PERSON_f1": 0.8838709677419355,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8404907975460123,
+      "eval_PERSON_precision": 0.85625,
       "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7945205479452054,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7435897435897436,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.15552844107151031,
-      "eval_overall_accuracy": 0.9581589958158996,
-      "eval_overall_f1": 0.8106666666666666,
-      "eval_overall_precision": 0.7619047619047619,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9939,
-      "eval_samples_per_second": 188.147,
-      "eval_steps_per_second": 3.018,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.14661699533462524,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8346883468834689,
+      "eval_overall_precision": 0.7958656330749354,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3911,
+      "eval_samples_per_second": 478.197,
+      "eval_steps_per_second": 7.672,
       "step": 3710
     },
     {
       "epoch": 36.0,
-      "grad_norm": 0.8952577710151672,
+      "grad_norm": 1.7864693403244019,
       "learning_rate": 3.2000000000000005e-05,
-      "loss": 0.0572,
+      "loss": 0.0562,
       "step": 3816
     },
     {
       "epoch": 36.0,
-      "eval_LOCATION_f1": 0.8157894736842105,
+      "eval_LOCATION_f1": 0.8026315789473685,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7272727272727272,
+      "eval_LOCATION_precision": 0.7721518987341772,
+      "eval_LOCATION_recall": 0.8356164383561644,
+      "eval_ORGANIZATION_f1": 0.7482993197278912,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6666666666666666,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8852459016393444,
+      "eval_ORGANIZATION_precision": 0.6707317073170732,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8838709677419355,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8709677419354839,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.85625,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7999999999999999,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8450704225352113,
+      "eval_QUANTITY_precision": 0.7222222222222222,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8219178082191781,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8108108108108109,
+      "eval_TIME_precision": 0.7692307692307693,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15078987181186676,
+      "eval_loss": 0.1466834545135498,
       "eval_overall_accuracy": 0.9625892197883338,
-      "eval_overall_f1": 0.8267394270122783,
-      "eval_overall_precision": 0.7931937172774869,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 1.0019,
-      "eval_samples_per_second": 186.652,
-      "eval_steps_per_second": 2.994,
+      "eval_overall_f1": 0.8273092369477911,
+      "eval_overall_precision": 0.7803030303030303,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3868,
+      "eval_samples_per_second": 483.41,
+      "eval_steps_per_second": 7.755,
       "step": 3816
     },
     {
       "epoch": 37.0,
-      "grad_norm": 1.4861292839050293,
+      "grad_norm": 0.5526124238967896,
       "learning_rate": 3.15e-05,
-      "loss": 0.0543,
+      "loss": 0.0548,
       "step": 3922
     },
     {
       "epoch": 37.0,
-      "eval_LOCATION_f1": 0.8181818181818182,
+      "eval_LOCATION_f1": 0.8051948051948052,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7777777777777778,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.6625766871165644,
+      "eval_LOCATION_precision": 0.7654320987654321,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.7346938775510204,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.5510204081632653,
+      "eval_ORGANIZATION_precision": 0.6585365853658537,
       "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8169014084507041,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8285714285714286,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7837837837837838,
+      "eval_TIME_precision": 0.8055555555555556,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.16852639615535736,
-      "eval_overall_accuracy": 0.9549593896135861,
-      "eval_overall_f1": 0.8042049934296978,
-      "eval_overall_precision": 0.7463414634146341,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9916,
-      "eval_samples_per_second": 188.593,
-      "eval_steps_per_second": 3.026,
+      "eval_loss": 0.15039868652820587,
+      "eval_overall_accuracy": 0.9613586020182131,
+      "eval_overall_f1": 0.8279569892473119,
+      "eval_overall_precision": 0.7837150127226463,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3874,
+      "eval_samples_per_second": 482.655,
+      "eval_steps_per_second": 7.743,
       "step": 3922
     },
     {
       "epoch": 38.0,
-      "grad_norm": 0.8263400197029114,
+      "grad_norm": 0.6235684156417847,
       "learning_rate": 3.1e-05,
-      "loss": 0.055,
+      "loss": 0.0528,
       "step": 4028
     },
     {
       "epoch": 38.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.8435374149659863,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7183098591549296,
+      "eval_LOCATION_precision": 0.8378378378378378,
+      "eval_LOCATION_recall": 0.8493150684931506,
+      "eval_ORGANIZATION_f1": 0.7313432835820897,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6623376623376623,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.7101449275362319,
+      "eval_ORGANIZATION_recall": 0.7538461538461538,
+      "eval_PERSON_f1": 0.8867313915857605,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.8616352201257862,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.75,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
+      "eval_QUANTITY_precision": 0.6857142857142857,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.861111111111111,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8157894736842105,
-      "eval_TIME_recall": 0.9117647058823529,
-      "eval_loss": 0.15474605560302734,
-      "eval_overall_accuracy": 0.9616047255722373,
-      "eval_overall_f1": 0.8217687074829932,
-      "eval_overall_precision": 0.7864583333333334,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.14205525815486908,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8342541436464089,
+      "eval_overall_precision": 0.8096514745308311,
       "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9962,
-      "eval_samples_per_second": 187.719,
-      "eval_steps_per_second": 3.012,
+      "eval_runtime": 0.3876,
+      "eval_samples_per_second": 482.503,
+      "eval_steps_per_second": 7.741,
       "step": 4028
     },
     {
       "epoch": 39.0,
-      "grad_norm": 0.7035608291625977,
+      "grad_norm": 0.9376897811889648,
       "learning_rate": 3.05e-05,
-      "loss": 0.051,
+      "loss": 0.0504,
       "step": 4134
     },
     {
       "epoch": 39.0,
-      "eval_LOCATION_f1": 0.832214765100671,
+      "eval_LOCATION_f1": 0.8496732026143791,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8157894736842105,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7517730496453899,
+      "eval_LOCATION_precision": 0.8125,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6973684210526315,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.71875,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8450704225352113,
+      "eval_QUANTITY_precision": 0.6571428571428571,
+      "eval_QUANTITY_recall": 0.7931034482758621,
+      "eval_TIME_f1": 0.8219178082191781,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8108108108108109,
+      "eval_TIME_precision": 0.7692307692307693,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.14542774856090546,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.8360655737704918,
-      "eval_overall_precision": 0.8031496062992126,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9872,
-      "eval_samples_per_second": 189.427,
-      "eval_steps_per_second": 3.039,
+      "eval_loss": 0.1570240557193756,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8355795148247979,
+      "eval_overall_precision": 0.7928388746803069,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3886,
+      "eval_samples_per_second": 481.159,
+      "eval_steps_per_second": 7.719,
       "step": 4134
     },
     {
       "epoch": 40.0,
-      "grad_norm": 0.368028461933136,
+      "grad_norm": 0.5903038382530212,
       "learning_rate": 3e-05,
       "loss": 0.053,
       "step": 4240
     },
     {
       "epoch": 40.0,
-      "eval_LOCATION_f1": 0.8133333333333332,
+      "eval_LOCATION_f1": 0.8366013071895425,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7922077922077922,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7066666666666667,
+      "eval_LOCATION_precision": 0.8,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7236842105263158,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6235294117647059,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8823529411764707,
+      "eval_ORGANIZATION_precision": 0.632183908045977,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8802588996763754,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8653846153846154,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.8553459119496856,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.7999999999999999,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7777777777777778,
-      "eval_TIME_recall": 0.8235294117647058,
-      "eval_loss": 0.15164527297019958,
-      "eval_overall_accuracy": 0.9598818606940684,
-      "eval_overall_f1": 0.8146143437077131,
-      "eval_overall_precision": 0.7757731958762887,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 1.0735,
-      "eval_samples_per_second": 174.198,
-      "eval_steps_per_second": 2.795,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.15270885825157166,
+      "eval_overall_accuracy": 0.9613586020182131,
+      "eval_overall_f1": 0.8304405874499333,
+      "eval_overall_precision": 0.7814070351758794,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3881,
+      "eval_samples_per_second": 481.816,
+      "eval_steps_per_second": 7.73,
       "step": 4240
     },
     {
       "epoch": 41.0,
-      "grad_norm": 0.21483705937862396,
+      "grad_norm": 0.2513221204280853,
       "learning_rate": 2.95e-05,
-      "loss": 0.0526,
+      "loss": 0.0469,
       "step": 4346
     },
     {
       "epoch": 41.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.8516129032258065,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7183098591549296,
-      "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6623376623376623,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_LOCATION_precision": 0.8048780487804879,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.75,
+      "eval_ORGANIZATION_number": 65,
+      "eval_ORGANIZATION_precision": 0.6835443037974683,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8774193548387097,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.85,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.6956521739130435,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
+      "eval_QUANTITY_precision": 0.6,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8055555555555555,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7631578947368421,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14916305243968964,
-      "eval_overall_accuracy": 0.9591434900319961,
-      "eval_overall_f1": 0.8092016238159675,
-      "eval_overall_precision": 0.770618556701031,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9975,
-      "eval_samples_per_second": 187.465,
-      "eval_steps_per_second": 3.007,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.15680240094661713,
+      "eval_overall_accuracy": 0.9593896135860202,
+      "eval_overall_f1": 0.8266666666666667,
+      "eval_overall_precision": 0.7769423558897243,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3878,
+      "eval_samples_per_second": 482.196,
+      "eval_steps_per_second": 7.736,
       "step": 4346
     },
     {
       "epoch": 42.0,
-      "grad_norm": 0.46089088916778564,
+      "grad_norm": 0.4710911512374878,
       "learning_rate": 2.9e-05,
-      "loss": 0.0497,
+      "loss": 0.0487,
       "step": 4452
     },
     {
       "epoch": 42.0,
-      "eval_LOCATION_f1": 0.8299319727891157,
+      "eval_LOCATION_f1": 0.8666666666666666,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8243243243243243,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7346938775510204,
+      "eval_LOCATION_precision": 0.8441558441558441,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7517730496453899,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6585365853658537,
-      "eval_ORGANIZATION_recall": 0.8307692307692308,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.6973684210526315,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7936507936507937,
+      "eval_PERSON_precision": 0.8670886075949367,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.696969696969697,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7352941176470589,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8169014084507041,
+      "eval_QUANTITY_precision": 0.6216216216216216,
+      "eval_QUANTITY_recall": 0.7931034482758621,
+      "eval_TIME_f1": 0.861111111111111,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7837837837837838,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.14921055734157562,
-      "eval_overall_accuracy": 0.963081466896382,
-      "eval_overall_f1": 0.8260869565217391,
-      "eval_overall_precision": 0.7896103896103897,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9983,
-      "eval_samples_per_second": 187.314,
-      "eval_steps_per_second": 3.005,
+      "eval_TIME_precision": 0.8157894736842105,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.14796730875968933,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8385345997286296,
+      "eval_overall_precision": 0.8005181347150259,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3855,
+      "eval_samples_per_second": 485.113,
+      "eval_steps_per_second": 7.783,
       "step": 4452
     },
     {
       "epoch": 43.0,
-      "grad_norm": 0.5246915221214294,
+      "grad_norm": 0.774330198764801,
       "learning_rate": 2.8499999999999998e-05,
-      "loss": 0.0499,
+      "loss": 0.0463,
       "step": 4558
     },
     {
       "epoch": 43.0,
-      "eval_LOCATION_f1": 0.7843137254901962,
+      "eval_LOCATION_f1": 0.8400000000000001,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.75,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7101449275362318,
+      "eval_LOCATION_precision": 0.8181818181818182,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7310344827586207,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6712328767123288,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6625,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8838709677419355,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.85625,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7076923076923076,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.6388888888888888,
+      "eval_QUANTITY_recall": 0.7931034482758621,
+      "eval_TIME_f1": 0.8450704225352113,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.8108108108108109,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15075072646141052,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8180574555403556,
-      "eval_overall_precision": 0.7868421052631579,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9913,
-      "eval_samples_per_second": 188.643,
-      "eval_steps_per_second": 3.026,
+      "eval_loss": 0.154840350151062,
+      "eval_overall_accuracy": 0.9618508491262614,
+      "eval_overall_f1": 0.8259109311740892,
+      "eval_overall_precision": 0.7846153846153846,
+      "eval_overall_recall": 0.8717948717948718,
+      "eval_runtime": 0.3848,
+      "eval_samples_per_second": 486.006,
+      "eval_steps_per_second": 7.797,
       "step": 4558
     },
     {
       "epoch": 44.0,
-      "grad_norm": 0.7562664151191711,
+      "grad_norm": 0.41449031233787537,
       "learning_rate": 2.8000000000000003e-05,
-      "loss": 0.0474,
+      "loss": 0.0459,
       "step": 4664
     },
     {
       "epoch": 44.0,
-      "eval_LOCATION_f1": 0.8400000000000001,
+      "eval_LOCATION_f1": 0.8456375838926175,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8181818181818182,
+      "eval_LOCATION_precision": 0.8289473684210527,
       "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7338129496402878,
+      "eval_ORGANIZATION_f1": 0.7692307692307693,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6891891891891891,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.887459807073955,
+      "eval_ORGANIZATION_precision": 0.7051282051282052,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8571428571428571,
-      "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7936507936507937,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7868852459016394,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7352941176470589,
-      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_QUANTITY_precision": 0.75,
+      "eval_QUANTITY_recall": 0.8275862068965517,
       "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
       "eval_TIME_precision": 0.8378378378378378,
       "eval_TIME_recall": 0.9117647058823529,
-      "eval_loss": 0.15113013982772827,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8392370572207084,
-      "eval_overall_precision": 0.804177545691906,
-      "eval_overall_recall": 0.8774928774928775,
-      "eval_runtime": 0.9996,
-      "eval_samples_per_second": 187.066,
-      "eval_steps_per_second": 3.001,
+      "eval_loss": 0.15145742893218994,
+      "eval_overall_accuracy": 0.9662810730986956,
+      "eval_overall_f1": 0.8512960436562074,
+      "eval_overall_precision": 0.8167539267015707,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3855,
+      "eval_samples_per_second": 485.134,
+      "eval_steps_per_second": 7.783,
       "step": 4664
     },
     {
       "epoch": 45.0,
-      "grad_norm": 3.280982494354248,
+      "grad_norm": 2.1948015689849854,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.0499,
+      "loss": 0.0454,
       "step": 4770
     },
     {
       "epoch": 45.0,
-      "eval_LOCATION_f1": 0.8456375838926175,
+      "eval_LOCATION_f1": 0.8609271523178808,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8289473684210527,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7375886524822695,
+      "eval_LOCATION_precision": 0.8333333333333334,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.702290076335878,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6842105263157895,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8896103896103896,
+      "eval_ORGANIZATION_precision": 0.696969696969697,
+      "eval_ORGANIZATION_recall": 0.7076923076923077,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8670886075949367,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8450704225352113,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8108108108108109,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15040434896945953,
-      "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8383561643835615,
-      "eval_overall_precision": 0.8073878627968337,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.995,
-      "eval_samples_per_second": 187.933,
-      "eval_steps_per_second": 3.015,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.1515069454908371,
+      "eval_overall_accuracy": 0.9657888259906473,
+      "eval_overall_f1": 0.850415512465374,
+      "eval_overall_precision": 0.8274932614555256,
+      "eval_overall_recall": 0.8746438746438746,
+      "eval_runtime": 0.3877,
+      "eval_samples_per_second": 482.27,
+      "eval_steps_per_second": 7.737,
       "step": 4770
     },
     {
       "epoch": 46.0,
-      "grad_norm": 0.43949922919273376,
+      "grad_norm": 0.26949745416641235,
       "learning_rate": 2.7000000000000002e-05,
-      "loss": 0.0477,
+      "loss": 0.0455,
       "step": 4876
     },
     {
       "epoch": 46.0,
-      "eval_LOCATION_f1": 0.8133333333333332,
+      "eval_LOCATION_f1": 0.8627450980392157,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7922077922077922,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.726027397260274,
+      "eval_LOCATION_precision": 0.825,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7681159420289856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.654320987654321,
+      "eval_ORGANIZATION_precision": 0.726027397260274,
       "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8896103896103896,
+      "eval_PERSON_f1": 0.8990228013029316,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8670886075949367,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8789808917197452,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8219178082191781,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7692307692307693,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15182223916053772,
-      "eval_overall_accuracy": 0.9625892197883338,
-      "eval_overall_f1": 0.8265582655826558,
-      "eval_overall_precision": 0.7881136950904393,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9979,
-      "eval_samples_per_second": 187.387,
-      "eval_steps_per_second": 3.006,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.14826659858226776,
+      "eval_overall_accuracy": 0.9672655673147921,
+      "eval_overall_f1": 0.8602739726027397,
+      "eval_overall_precision": 0.8284960422163589,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3876,
+      "eval_samples_per_second": 482.517,
+      "eval_steps_per_second": 7.741,
       "step": 4876
     },
     {
       "epoch": 47.0,
-      "grad_norm": 1.4698158502578735,
+      "grad_norm": 1.137783408164978,
       "learning_rate": 2.6500000000000004e-05,
-      "loss": 0.0463,
+      "loss": 0.0424,
       "step": 4982
     },
     {
       "epoch": 47.0,
-      "eval_LOCATION_f1": 0.8133333333333332,
+      "eval_LOCATION_f1": 0.8684210526315789,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7922077922077922,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.75,
+      "eval_LOCATION_precision": 0.8354430379746836,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7534246575342466,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6551724137931034,
-      "eval_ORGANIZATION_recall": 0.8769230769230769,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.6790123456790124,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8938906752411575,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8633540372670807,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8055555555555555,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7631578947368421,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.15979216992855072,
-      "eval_overall_accuracy": 0.9596357371400444,
-      "eval_overall_f1": 0.822429906542056,
-      "eval_overall_precision": 0.7738693467336684,
-      "eval_overall_recall": 0.8774928774928775,
-      "eval_runtime": 0.9931,
-      "eval_samples_per_second": 188.295,
-      "eval_steps_per_second": 3.021,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.15139973163604736,
+      "eval_overall_accuracy": 0.9655427024366232,
+      "eval_overall_f1": 0.8540540540540541,
+      "eval_overall_precision": 0.8123393316195373,
+      "eval_overall_recall": 0.9002849002849003,
+      "eval_runtime": 0.3866,
+      "eval_samples_per_second": 483.755,
+      "eval_steps_per_second": 7.761,
       "step": 4982
     },
     {
       "epoch": 48.0,
-      "grad_norm": 0.8823111057281494,
+      "grad_norm": 2.368671178817749,
       "learning_rate": 2.6000000000000002e-05,
-      "loss": 0.0458,
+      "loss": 0.0444,
       "step": 5088
     },
     {
       "epoch": 48.0,
-      "eval_LOCATION_f1": 0.8076923076923077,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7590361445783133,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7123287671232877,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7611940298507462,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6419753086419753,
-      "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.7391304347826086,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7272727272727273,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
+      "eval_QUANTITY_precision": 0.6486486486486487,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8333333333333333,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7894736842105263,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16684772074222565,
-      "eval_overall_accuracy": 0.9598818606940684,
-      "eval_overall_f1": 0.8214765100671142,
-      "eval_overall_precision": 0.7766497461928934,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9966,
-      "eval_samples_per_second": 187.646,
-      "eval_steps_per_second": 3.01,
+      "eval_TIME_precision": 0.7631578947368421,
+      "eval_TIME_recall": 0.8529411764705882,
+      "eval_loss": 0.15529875457286835,
+      "eval_overall_accuracy": 0.965296578882599,
+      "eval_overall_f1": 0.8415300546448087,
+      "eval_overall_precision": 0.8083989501312336,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3865,
+      "eval_samples_per_second": 483.867,
+      "eval_steps_per_second": 7.763,
       "step": 5088
     },
     {
       "epoch": 49.0,
-      "grad_norm": 0.5812783241271973,
+      "grad_norm": 0.1998334676027298,
       "learning_rate": 2.5500000000000003e-05,
-      "loss": 0.0428,
+      "loss": 0.0427,
       "step": 5194
     },
     {
       "epoch": 49.0,
-      "eval_LOCATION_f1": 0.8076923076923077,
+      "eval_LOCATION_f1": 0.88,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7590361445783133,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.726027397260274,
+      "eval_LOCATION_precision": 0.8571428571428571,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7972027972027972,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.654320987654321,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8774193548387097,
+      "eval_ORGANIZATION_precision": 0.7307692307692307,
+      "eval_ORGANIZATION_recall": 0.8769230769230769,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8219178082191781,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7692307692307693,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16307547688484192,
-      "eval_overall_accuracy": 0.9596357371400444,
-      "eval_overall_f1": 0.8192771084337349,
-      "eval_overall_precision": 0.7727272727272727,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9928,
-      "eval_samples_per_second": 188.354,
-      "eval_steps_per_second": 3.022,
+      "eval_loss": 0.15635553002357483,
+      "eval_overall_accuracy": 0.965296578882599,
+      "eval_overall_f1": 0.8602442333785618,
+      "eval_overall_precision": 0.8212435233160622,
+      "eval_overall_recall": 0.9031339031339032,
+      "eval_runtime": 0.3869,
+      "eval_samples_per_second": 483.349,
+      "eval_steps_per_second": 7.754,
       "step": 5194
     },
     {
       "epoch": 50.0,
-      "grad_norm": 0.9438253045082092,
+      "grad_norm": 0.9808441996574402,
       "learning_rate": 2.5e-05,
-      "loss": 0.0437,
+      "loss": 0.041,
       "step": 5300
     },
     {
       "epoch": 50.0,
-      "eval_LOCATION_f1": 0.7792207792207793,
+      "eval_LOCATION_f1": 0.855263157894737,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7407407407407407,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7272727272727273,
+      "eval_LOCATION_precision": 0.8227848101265823,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7857142857142856,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7164179104477612,
-      "eval_ORGANIZATION_recall": 0.7384615384615385,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.7333333333333333,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9032258064516129,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.875,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8857142857142858,
+      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8611111111111112,
-      "eval_TIME_recall": 0.9117647058823529,
-      "eval_loss": 0.15793509781360626,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8225584594222833,
-      "eval_overall_precision": 0.7952127659574468,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9991,
-      "eval_samples_per_second": 187.177,
-      "eval_steps_per_second": 3.003,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.15144038200378418,
+      "eval_overall_accuracy": 0.9660349495446714,
+      "eval_overall_f1": 0.8571428571428573,
+      "eval_overall_precision": 0.8203125,
+      "eval_overall_recall": 0.8974358974358975,
+      "eval_runtime": 0.3879,
+      "eval_samples_per_second": 482.064,
+      "eval_steps_per_second": 7.734,
       "step": 5300
     },
     {
       "epoch": 51.0,
-      "grad_norm": 0.8421459794044495,
+      "grad_norm": 0.43952253460884094,
       "learning_rate": 2.45e-05,
-      "loss": 0.0436,
+      "loss": 0.0416,
       "step": 5406
     },
     {
       "epoch": 51.0,
-      "eval_LOCATION_f1": 0.8026315789473685,
+      "eval_LOCATION_f1": 0.858974358974359,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7721518987341772,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7092198581560285,
+      "eval_LOCATION_precision": 0.8072289156626506,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7246376811594203,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6578947368421053,
+      "eval_ORGANIZATION_precision": 0.684931506849315,
       "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_PERSON_f1": 0.8817891373801918,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8466257668711656,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15914805233478546,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.8240109140518418,
-      "eval_overall_precision": 0.7905759162303665,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9849,
-      "eval_samples_per_second": 189.872,
-      "eval_steps_per_second": 3.046,
+      "eval_loss": 0.1590029001235962,
+      "eval_overall_accuracy": 0.9650504553285749,
+      "eval_overall_f1": 0.8405405405405405,
+      "eval_overall_precision": 0.7994858611825193,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3863,
+      "eval_samples_per_second": 484.101,
+      "eval_steps_per_second": 7.766,
       "step": 5406
     },
     {
       "epoch": 52.0,
-      "grad_norm": 0.6999145150184631,
+      "grad_norm": 0.5576232075691223,
       "learning_rate": 2.4e-05,
-      "loss": 0.0403,
+      "loss": 0.0398,
       "step": 5512
     },
     {
       "epoch": 52.0,
-      "eval_LOCATION_f1": 0.8181818181818182,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7777777777777778,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7066666666666667,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7516778523489933,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6235294117647059,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.8737864077669903,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
+      "eval_PERSON_f1": 0.8910256410256411,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8490566037735849,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.8580246913580247,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
       "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
       "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1684279888868332,
-      "eval_overall_accuracy": 0.9593896135860202,
-      "eval_overall_f1": 0.8155080213903743,
-      "eval_overall_precision": 0.7682619647355163,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9863,
-      "eval_samples_per_second": 189.596,
-      "eval_steps_per_second": 3.042,
+      "eval_loss": 0.16141797602176666,
+      "eval_overall_accuracy": 0.9648043317745508,
+      "eval_overall_f1": 0.8498659517426275,
+      "eval_overall_precision": 0.8025316455696202,
+      "eval_overall_recall": 0.9031339031339032,
+      "eval_runtime": 0.3866,
+      "eval_samples_per_second": 483.708,
+      "eval_steps_per_second": 7.76,
       "step": 5512
     },
     {
       "epoch": 53.0,
-      "grad_norm": 0.8465991616249084,
+      "grad_norm": 0.973469078540802,
       "learning_rate": 2.35e-05,
-      "loss": 0.0447,
+      "loss": 0.0404,
       "step": 5618
     },
     {
       "epoch": 53.0,
-      "eval_LOCATION_f1": 0.8400000000000001,
+      "eval_LOCATION_f1": 0.8684210526315789,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8181818181818182,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7361111111111112,
+      "eval_LOCATION_precision": 0.8354430379746836,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6708860759493671,
-      "eval_ORGANIZATION_recall": 0.8153846153846154,
-      "eval_PERSON_f1": 0.896774193548387,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9067524115755627,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.86875,
-      "eval_PERSON_recall": 0.9266666666666666,
-      "eval_QUANTITY_f1": 0.7575757575757576,
+      "eval_PERSON_precision": 0.8757763975155279,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6756756756756757,
+      "eval_QUANTITY_precision": 0.7142857142857143,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8219178082191781,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7692307692307693,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16614499688148499,
-      "eval_overall_accuracy": 0.9606202313561408,
-      "eval_overall_f1": 0.8344549125168237,
-      "eval_overall_precision": 0.7908163265306123,
-      "eval_overall_recall": 0.8831908831908832,
-      "eval_runtime": 0.9842,
-      "eval_samples_per_second": 190.011,
-      "eval_steps_per_second": 3.048,
+      "eval_loss": 0.1627216637134552,
+      "eval_overall_accuracy": 0.963081466896382,
+      "eval_overall_f1": 0.8517520215633424,
+      "eval_overall_precision": 0.8081841432225064,
+      "eval_overall_recall": 0.9002849002849003,
+      "eval_runtime": 0.3859,
+      "eval_samples_per_second": 484.521,
+      "eval_steps_per_second": 7.773,
       "step": 5618
     },
     {
       "epoch": 54.0,
-      "grad_norm": 0.7064285278320312,
+      "grad_norm": 0.923568606376648,
       "learning_rate": 2.3000000000000003e-05,
-      "loss": 0.0411,
+      "loss": 0.0405,
       "step": 5724
     },
     {
       "epoch": 54.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.8684210526315789,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.7183098591549296,
+      "eval_LOCATION_precision": 0.8354430379746836,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7361111111111112,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6623376623376623,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8925081433224756,
+      "eval_ORGANIZATION_precision": 0.6708860759493671,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8726114649681529,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16279193758964539,
-      "eval_overall_accuracy": 0.9603741078021166,
-      "eval_overall_f1": 0.826265389876881,
-      "eval_overall_precision": 0.7947368421052632,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9899,
-      "eval_samples_per_second": 188.907,
-      "eval_steps_per_second": 3.031,
+      "eval_loss": 0.15863551199436188,
+      "eval_overall_accuracy": 0.9660349495446714,
+      "eval_overall_f1": 0.8497970230040596,
+      "eval_overall_precision": 0.8092783505154639,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3852,
+      "eval_samples_per_second": 485.517,
+      "eval_steps_per_second": 7.789,
       "step": 5724
     },
     {
       "epoch": 55.0,
-      "grad_norm": 0.41567376255989075,
+      "grad_norm": 0.5156478881835938,
       "learning_rate": 2.25e-05,
-      "loss": 0.039,
+      "loss": 0.037,
       "step": 5830
     },
     {
       "epoch": 55.0,
-      "eval_LOCATION_f1": 0.794701986754967,
+      "eval_LOCATION_f1": 0.8441558441558442,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7692307692307693,
-      "eval_LOCATION_recall": 0.821917808219178,
-      "eval_ORGANIZATION_f1": 0.6950354609929077,
+      "eval_LOCATION_precision": 0.8024691358024691,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7338129496402878,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6447368421052632,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6891891891891891,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8810289389067525,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.75,
+      "eval_PERSON_precision": 0.8509316770186336,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.7692307692307692,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6857142857142857,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8450704225352113,
+      "eval_QUANTITY_precision": 0.6944444444444444,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8108108108108109,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16604018211364746,
-      "eval_overall_accuracy": 0.9581589958158996,
-      "eval_overall_f1": 0.8147138964577657,
-      "eval_overall_precision": 0.7806788511749347,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9924,
-      "eval_samples_per_second": 188.441,
-      "eval_steps_per_second": 3.023,
+      "eval_loss": 0.16379383206367493,
+      "eval_overall_accuracy": 0.9618508491262614,
+      "eval_overall_f1": 0.8313090418353576,
+      "eval_overall_precision": 0.7897435897435897,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3848,
+      "eval_samples_per_second": 485.964,
+      "eval_steps_per_second": 7.796,
       "step": 5830
     },
     {
       "epoch": 56.0,
-      "grad_norm": 3.736433744430542,
+      "grad_norm": 1.4201856851577759,
       "learning_rate": 2.2000000000000003e-05,
-      "loss": 0.0416,
+      "loss": 0.0367,
       "step": 5936
     },
     {
       "epoch": 56.0,
-      "eval_LOCATION_f1": 0.8157894736842105,
+      "eval_LOCATION_f1": 0.8513513513513513,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7848101265822784,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7246376811594203,
+      "eval_LOCATION_precision": 0.84,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7391304347826088,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.684931506849315,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8932038834951457,
+      "eval_ORGANIZATION_precision": 0.6986301369863014,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8846153846153846,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_precision": 0.8518518518518519,
       "eval_PERSON_recall": 0.92,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.861111111111111,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1560632586479187,
-      "eval_overall_accuracy": 0.9628353433423579,
-      "eval_overall_f1": 0.832876712328767,
-      "eval_overall_precision": 0.8021108179419525,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9995,
-      "eval_samples_per_second": 187.088,
-      "eval_steps_per_second": 3.001,
+      "eval_TIME_precision": 0.8157894736842105,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.15849658846855164,
+      "eval_overall_accuracy": 0.9648043317745508,
+      "eval_overall_f1": 0.840381991814461,
+      "eval_overall_precision": 0.806282722513089,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3864,
+      "eval_samples_per_second": 483.981,
+      "eval_steps_per_second": 7.764,
       "step": 5936
     },
     {
       "epoch": 57.0,
-      "grad_norm": 0.16218724846839905,
+      "grad_norm": 0.06779062002897263,
       "learning_rate": 2.15e-05,
-      "loss": 0.0386,
+      "loss": 0.0387,
       "step": 6042
     },
     {
       "epoch": 57.0,
-      "eval_LOCATION_f1": 0.8051948051948052,
+      "eval_LOCATION_f1": 0.8571428571428572,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7654320987654321,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7313432835820897,
+      "eval_LOCATION_precision": 0.8513513513513513,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7482014388489208,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7101449275362319,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.7027027027027027,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7812500000000001,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7142857142857143,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8333333333333333,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7894736842105263,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1614658534526825,
-      "eval_overall_accuracy": 0.9616047255722373,
-      "eval_overall_f1": 0.8240109140518418,
-      "eval_overall_precision": 0.7905759162303665,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9919,
-      "eval_samples_per_second": 188.532,
-      "eval_steps_per_second": 3.025,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.15448686480522156,
+      "eval_overall_accuracy": 0.9662810730986956,
+      "eval_overall_f1": 0.8516483516483517,
+      "eval_overall_precision": 0.8222811671087533,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3879,
+      "eval_samples_per_second": 482.058,
+      "eval_steps_per_second": 7.734,
       "step": 6042
     },
     {
       "epoch": 58.0,
-      "grad_norm": 0.9404721856117249,
+      "grad_norm": 0.30816060304641724,
       "learning_rate": 2.1e-05,
-      "loss": 0.0395,
+      "loss": 0.0371,
       "step": 6148
     },
     {
       "epoch": 58.0,
-      "eval_LOCATION_f1": 0.8076923076923077,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7590361445783133,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7313432835820897,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7101449275362319,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8867313915857605,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8774193548387097,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8616352201257862,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7619047619047619,
+      "eval_PERSON_precision": 0.85,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7058823529411765,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8450704225352113,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8108108108108109,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1566106230020523,
-      "eval_overall_accuracy": 0.9625892197883338,
-      "eval_overall_f1": 0.8290013679890562,
-      "eval_overall_precision": 0.7973684210526316,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9978,
-      "eval_samples_per_second": 187.413,
-      "eval_steps_per_second": 3.007,
+      "eval_loss": 0.16250932216644287,
+      "eval_overall_accuracy": 0.963081466896382,
+      "eval_overall_f1": 0.8358208955223881,
+      "eval_overall_precision": 0.7979274611398963,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3884,
+      "eval_samples_per_second": 481.478,
+      "eval_steps_per_second": 7.724,
       "step": 6148
     },
     {
       "epoch": 59.0,
-      "grad_norm": 1.2607734203338623,
+      "grad_norm": 0.8676481246948242,
       "learning_rate": 2.05e-05,
-      "loss": 0.0376,
+      "loss": 0.0354,
       "step": 6254
     },
     {
       "epoch": 59.0,
-      "eval_LOCATION_f1": 0.8258064516129032,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7804878048780488,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7218045112781954,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7391304347826088,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7058823529411765,
-      "eval_ORGANIZATION_recall": 0.7384615384615385,
-      "eval_PERSON_f1": 0.8701298701298702,
+      "eval_ORGANIZATION_precision": 0.6986301369863014,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8810289389067525,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8481012658227848,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8509316770186336,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16003942489624023,
-      "eval_overall_accuracy": 0.9633275904504061,
-      "eval_overall_f1": 0.8253094910591472,
-      "eval_overall_precision": 0.7978723404255319,
-      "eval_overall_recall": 0.8547008547008547,
-      "eval_runtime": 1.001,
-      "eval_samples_per_second": 186.811,
-      "eval_steps_per_second": 2.997,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.16529938578605652,
+      "eval_overall_accuracy": 0.965296578882599,
+      "eval_overall_f1": 0.8474114441416893,
+      "eval_overall_precision": 0.8120104438642297,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3866,
+      "eval_samples_per_second": 483.696,
+      "eval_steps_per_second": 7.76,
       "step": 6254
     },
     {
       "epoch": 60.0,
-      "grad_norm": 0.5839110612869263,
+      "grad_norm": 0.16957683861255646,
       "learning_rate": 2e-05,
-      "loss": 0.0364,
+      "loss": 0.0363,
       "step": 6360
     },
     {
       "epoch": 60.0,
-      "eval_LOCATION_f1": 0.8378378378378378,
+      "eval_LOCATION_f1": 0.88,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8266666666666667,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7205882352941175,
+      "eval_LOCATION_precision": 0.8571428571428571,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7375886524822695,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6901408450704225,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8758169934640523,
+      "eval_ORGANIZATION_precision": 0.6842105263157895,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8846153846153846,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8589743589743589,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8518518518518519,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.15972208976745605,
-      "eval_overall_accuracy": 0.9633275904504061,
-      "eval_overall_f1": 0.8294036061026352,
-      "eval_overall_precision": 0.8081081081081081,
-      "eval_overall_recall": 0.8518518518518519,
-      "eval_runtime": 0.9927,
-      "eval_samples_per_second": 188.383,
-      "eval_steps_per_second": 3.022,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.16163839399814606,
+      "eval_overall_accuracy": 0.9655427024366232,
+      "eval_overall_f1": 0.8528610354223433,
+      "eval_overall_precision": 0.8172323759791122,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3856,
+      "eval_samples_per_second": 484.992,
+      "eval_steps_per_second": 7.781,
       "step": 6360
     },
     {
       "epoch": 61.0,
-      "grad_norm": 0.9700207710266113,
+      "grad_norm": 0.5314693450927734,
       "learning_rate": 1.9500000000000003e-05,
-      "loss": 0.0362,
+      "loss": 0.0333,
       "step": 6466
     },
     {
       "epoch": 61.0,
-      "eval_LOCATION_f1": 0.8513513513513513,
+      "eval_LOCATION_f1": 0.847682119205298,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.84,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7132867132867133,
+      "eval_LOCATION_precision": 0.8205128205128205,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7027027027027027,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6538461538461539,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8896103896103896,
+      "eval_ORGANIZATION_precision": 0.6265060240963856,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8670886075949367,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8571428571428571,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16418209671974182,
-      "eval_overall_accuracy": 0.9618508491262614,
-      "eval_overall_f1": 0.8344733242134063,
-      "eval_overall_precision": 0.8026315789473685,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9901,
-      "eval_samples_per_second": 188.877,
-      "eval_steps_per_second": 3.03,
+      "eval_loss": 0.17665402591228485,
+      "eval_overall_accuracy": 0.9623430962343096,
+      "eval_overall_f1": 0.8322147651006713,
+      "eval_overall_precision": 0.7868020304568528,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3885,
+      "eval_samples_per_second": 481.387,
+      "eval_steps_per_second": 7.723,
       "step": 6466
     },
     {
       "epoch": 62.0,
-      "grad_norm": 0.7778552770614624,
+      "grad_norm": 0.8911972641944885,
       "learning_rate": 1.9e-05,
-      "loss": 0.0358,
+      "loss": 0.0324,
       "step": 6572
     },
     {
       "epoch": 62.0,
-      "eval_LOCATION_f1": 0.8366013071895425,
+      "eval_LOCATION_f1": 0.8456375838926175,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7338129496402878,
+      "eval_LOCATION_precision": 0.8289473684210527,
+      "eval_LOCATION_recall": 0.863013698630137,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6891891891891891,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8888888888888887,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9090909090909091,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8717948717948718,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7692307692307692,
+      "eval_PERSON_precision": 0.8860759493670886,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.6944444444444444,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8450704225352113,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.8108108108108109,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1708838939666748,
-      "eval_overall_accuracy": 0.9613586020182131,
-      "eval_overall_f1": 0.834924965893588,
-      "eval_overall_precision": 0.8010471204188482,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 1.08,
-      "eval_samples_per_second": 173.149,
-      "eval_steps_per_second": 2.778,
+      "eval_loss": 0.16479013860225677,
+      "eval_overall_accuracy": 0.965296578882599,
+      "eval_overall_f1": 0.8524590163934426,
+      "eval_overall_precision": 0.8188976377952756,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3895,
+      "eval_samples_per_second": 480.151,
+      "eval_steps_per_second": 7.703,
       "step": 6572
     },
     {
       "epoch": 63.0,
-      "grad_norm": 2.994206428527832,
+      "grad_norm": 1.928049087524414,
       "learning_rate": 1.85e-05,
-      "loss": 0.034,
+      "loss": 0.0364,
       "step": 6678
     },
     {
       "epoch": 63.0,
-      "eval_LOCATION_f1": 0.8571428571428572,
+      "eval_LOCATION_f1": 0.8648648648648648,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8513513513513513,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7338129496402878,
+      "eval_LOCATION_precision": 0.8533333333333334,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7586206896551724,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6891891891891891,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8729641693811074,
+      "eval_ORGANIZATION_precision": 0.6875,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8535031847133758,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.164272740483284,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8381742738589213,
-      "eval_overall_precision": 0.8145161290322581,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9945,
-      "eval_samples_per_second": 188.026,
-      "eval_steps_per_second": 3.016,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.16289159655570984,
+      "eval_overall_accuracy": 0.9650504553285749,
+      "eval_overall_f1": 0.8555858310626703,
+      "eval_overall_precision": 0.8198433420365535,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3869,
+      "eval_samples_per_second": 483.389,
+      "eval_steps_per_second": 7.755,
       "step": 6678
     },
     {
       "epoch": 64.0,
-      "grad_norm": 0.950914740562439,
+      "grad_norm": 0.33195340633392334,
       "learning_rate": 1.8e-05,
-      "loss": 0.0353,
+      "loss": 0.034,
       "step": 6784
     },
     {
       "epoch": 64.0,
-      "eval_LOCATION_f1": 0.8441558441558442,
+      "eval_LOCATION_f1": 0.8571428571428571,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8024691358024691,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7272727272727273,
+      "eval_LOCATION_precision": 0.8148148148148148,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.6616541353383458,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7164179104477612,
-      "eval_ORGANIZATION_recall": 0.7384615384615385,
-      "eval_PERSON_f1": 0.8758169934640523,
+      "eval_ORGANIZATION_precision": 0.6470588235294118,
+      "eval_ORGANIZATION_recall": 0.676923076923077,
+      "eval_PERSON_f1": 0.8802588996763754,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8589743589743589,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8553459119496856,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8857142857142858,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
-      "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16246764361858368,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8337950138504155,
-      "eval_overall_precision": 0.8113207547169812,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 0.994,
-      "eval_samples_per_second": 188.133,
-      "eval_steps_per_second": 3.018,
+      "eval_TIME_precision": 0.8611111111111112,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.1654130518436432,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8324175824175825,
+      "eval_overall_precision": 0.8037135278514589,
+      "eval_overall_recall": 0.8632478632478633,
+      "eval_runtime": 0.3879,
+      "eval_samples_per_second": 482.096,
+      "eval_steps_per_second": 7.734,
       "step": 6784
     },
     {
       "epoch": 65.0,
-      "grad_norm": 0.5340273976325989,
+      "grad_norm": 0.7465774416923523,
       "learning_rate": 1.75e-05,
-      "loss": 0.0335,
+      "loss": 0.0325,
       "step": 6890
     },
     {
       "epoch": 65.0,
-      "eval_LOCATION_f1": 0.8441558441558442,
+      "eval_LOCATION_f1": 0.8724832214765101,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8024691358024691,
+      "eval_LOCATION_precision": 0.8552631578947368,
       "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7407407407407408,
+      "eval_ORGANIZATION_f1": 0.7428571428571429,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7142857142857143,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.6933333333333334,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8571428571428571,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8219178082191781,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.7317073170731707,
+      "eval_TIME_precision": 0.7692307692307693,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17117133736610413,
-      "eval_overall_accuracy": 0.9628353433423579,
-      "eval_overall_f1": 0.8283378746594006,
-      "eval_overall_precision": 0.793733681462141,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9791,
-      "eval_samples_per_second": 190.991,
-      "eval_steps_per_second": 3.064,
+      "eval_loss": 0.1645297259092331,
+      "eval_overall_accuracy": 0.9665271966527197,
+      "eval_overall_f1": 0.8435374149659863,
+      "eval_overall_precision": 0.8072916666666666,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3875,
+      "eval_samples_per_second": 482.599,
+      "eval_steps_per_second": 7.742,
       "step": 6890
     },
     {
       "epoch": 66.0,
-      "grad_norm": 0.4983876049518585,
+      "grad_norm": 2.1847379207611084,
       "learning_rate": 1.7000000000000003e-05,
-      "loss": 0.0344,
+      "loss": 0.0322,
       "step": 6996
     },
     {
       "epoch": 66.0,
-      "eval_LOCATION_f1": 0.8344370860927152,
-      "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8076923076923077,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7428571428571429,
+      "eval_LOCATION_f1": 0.8571428571428571,
+      "eval_LOCATION_number": 73,
+      "eval_LOCATION_precision": 0.8148148148148148,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7272727272727272,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6933333333333334,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
       "eval_ORGANIZATION_recall": 0.8,
-      "eval_PERSON_f1": 0.8838709677419355,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.85625,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
       "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
       "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.16945300996303558,
-      "eval_overall_accuracy": 0.9623430962343096,
-      "eval_overall_f1": 0.8326530612244898,
-      "eval_overall_precision": 0.796875,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9853,
-      "eval_samples_per_second": 189.785,
-      "eval_steps_per_second": 3.045,
+      "eval_loss": 0.1727007031440735,
+      "eval_overall_accuracy": 0.9645582082205267,
+      "eval_overall_f1": 0.845945945945946,
+      "eval_overall_precision": 0.8046272493573264,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3869,
+      "eval_samples_per_second": 483.358,
+      "eval_steps_per_second": 7.754,
       "step": 6996
     },
     {
       "epoch": 67.0,
-      "grad_norm": 0.8631144762039185,
+      "grad_norm": 0.15823009610176086,
       "learning_rate": 1.65e-05,
-      "loss": 0.0333,
+      "loss": 0.0314,
       "step": 7102
     },
     {
       "epoch": 67.0,
-      "eval_LOCATION_f1": 0.8441558441558442,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8024691358024691,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7246376811594203,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7014925373134329,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.684931506849315,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.6811594202898551,
+      "eval_ORGANIZATION_recall": 0.7230769230769231,
+      "eval_PERSON_f1": 0.896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8064516129032258,
+      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_precision": 0.7352941176470589,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.7945205479452054,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
+      "eval_TIME_precision": 0.7435897435897436,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1660642921924591,
-      "eval_overall_accuracy": 0.9633275904504061,
-      "eval_overall_f1": 0.8317373461012312,
-      "eval_overall_precision": 0.8,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9933,
-      "eval_samples_per_second": 188.27,
-      "eval_steps_per_second": 3.02,
+      "eval_loss": 0.16559243202209473,
+      "eval_overall_accuracy": 0.9643120846665025,
+      "eval_overall_f1": 0.8367626886145405,
+      "eval_overall_precision": 0.8068783068783069,
+      "eval_overall_recall": 0.8689458689458689,
+      "eval_runtime": 0.3869,
+      "eval_samples_per_second": 483.332,
+      "eval_steps_per_second": 7.754,
       "step": 7102
     },
     {
       "epoch": 68.0,
-      "grad_norm": 0.16754050552845,
+      "grad_norm": 0.8872466087341309,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.0344,
+      "loss": 0.0313,
       "step": 7208
     },
     {
       "epoch": 68.0,
-      "eval_LOCATION_f1": 0.8441558441558442,
+      "eval_LOCATION_f1": 0.8724832214765101,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8024691358024691,
+      "eval_LOCATION_precision": 0.8552631578947368,
       "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7313432835820897,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7101449275362319,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.8831168831168831,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8064516129032258,
+      "eval_PERSON_precision": 0.8607594936708861,
+      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_precision": 0.7352941176470589,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8055555555555555,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
+      "eval_TIME_precision": 0.7631578947368421,
       "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1679677516222,
-      "eval_overall_accuracy": 0.9638198375584544,
-      "eval_overall_f1": 0.8347107438016529,
-      "eval_overall_precision": 0.808,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9969,
-      "eval_samples_per_second": 187.58,
-      "eval_steps_per_second": 3.009,
+      "eval_loss": 0.17109757661819458,
+      "eval_overall_accuracy": 0.9643120846665025,
+      "eval_overall_f1": 0.8392370572207084,
+      "eval_overall_precision": 0.804177545691906,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3881,
+      "eval_samples_per_second": 481.831,
+      "eval_steps_per_second": 7.73,
       "step": 7208
     },
     {
       "epoch": 69.0,
-      "grad_norm": 0.08073262125253677,
+      "grad_norm": 0.13313980400562286,
       "learning_rate": 1.55e-05,
-      "loss": 0.0297,
+      "loss": 0.0319,
       "step": 7314
     },
     {
       "epoch": 69.0,
-      "eval_LOCATION_f1": 0.8366013071895425,
+      "eval_LOCATION_f1": 0.8666666666666666,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7222222222222222,
+      "eval_LOCATION_precision": 0.8441558441558441,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7671232876712328,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6582278481012658,
-      "eval_ORGANIZATION_recall": 0.8,
+      "eval_ORGANIZATION_precision": 0.691358024691358,
+      "eval_ORGANIZATION_recall": 0.8615384615384616,
       "eval_PERSON_f1": 0.8802588996763754,
       "eval_PERSON_number": 150,
       "eval_PERSON_precision": 0.8553459119496856,
       "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7936507936507937,
+      "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_precision": 0.78125,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17325514554977417,
-      "eval_overall_accuracy": 0.9618508491262614,
-      "eval_overall_f1": 0.8292682926829267,
-      "eval_overall_precision": 0.7906976744186046,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9946,
-      "eval_samples_per_second": 188.01,
-      "eval_steps_per_second": 3.016,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.17586369812488556,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8455284552845528,
+      "eval_overall_precision": 0.8062015503875969,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3858,
+      "eval_samples_per_second": 484.678,
+      "eval_steps_per_second": 7.776,
       "step": 7314
     },
     {
       "epoch": 70.0,
-      "grad_norm": 0.45353350043296814,
+      "grad_norm": 0.6439195871353149,
       "learning_rate": 1.5e-05,
-      "loss": 0.0304,
+      "loss": 0.0303,
       "step": 7420
     },
     {
       "epoch": 70.0,
-      "eval_LOCATION_f1": 0.8421052631578948,
+      "eval_LOCATION_f1": 0.8724832214765101,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.810126582278481,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.75,
+      "eval_LOCATION_precision": 0.8552631578947368,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7428571428571429,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7183098591549296,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8729641693811074,
+      "eval_ORGANIZATION_precision": 0.6933333333333334,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8535031847133758,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_precision": 0.7352941176470589,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8219178082191781,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17020219564437866,
-      "eval_overall_accuracy": 0.9638198375584544,
-      "eval_overall_f1": 0.8358620689655173,
-      "eval_overall_precision": 0.8101604278074866,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9973,
-      "eval_samples_per_second": 187.51,
-      "eval_steps_per_second": 3.008,
+      "eval_TIME_precision": 0.7692307692307693,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.1740642637014389,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8462585034013606,
+      "eval_overall_precision": 0.8098958333333334,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3854,
+      "eval_samples_per_second": 485.205,
+      "eval_steps_per_second": 7.784,
       "step": 7420
     },
     {
       "epoch": 71.0,
-      "grad_norm": 1.6932181119918823,
+      "grad_norm": 1.417320728302002,
       "learning_rate": 1.45e-05,
-      "loss": 0.0329,
+      "loss": 0.0299,
       "step": 7526
     },
     {
       "epoch": 71.0,
-      "eval_LOCATION_f1": 0.8333333333333334,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7831325301204819,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7050359712230215,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6621621621621622,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8701298701298702,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8481012658227848,
-      "eval_PERSON_recall": 0.8933333333333333,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
       "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
       "eval_QUANTITY_precision": 0.78125,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8732394366197184,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17471523582935333,
-      "eval_overall_accuracy": 0.963081466896382,
-      "eval_overall_f1": 0.8240109140518418,
-      "eval_overall_precision": 0.7905759162303665,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9916,
-      "eval_samples_per_second": 188.583,
-      "eval_steps_per_second": 3.025,
+      "eval_TIME_precision": 0.8378378378378378,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.17045490443706512,
+      "eval_overall_accuracy": 0.9650504553285749,
+      "eval_overall_f1": 0.8571428571428573,
+      "eval_overall_precision": 0.8203125,
+      "eval_overall_recall": 0.8974358974358975,
+      "eval_runtime": 0.3874,
+      "eval_samples_per_second": 482.652,
+      "eval_steps_per_second": 7.743,
       "step": 7526
     },
     {
       "epoch": 72.0,
-      "grad_norm": 0.46643197536468506,
+      "grad_norm": 0.37806442379951477,
       "learning_rate": 1.4000000000000001e-05,
-      "loss": 0.0319,
+      "loss": 0.0289,
       "step": 7632
     },
     {
       "epoch": 72.0,
-      "eval_LOCATION_f1": 0.8299319727891157,
+      "eval_LOCATION_f1": 0.855263157894737,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8243243243243243,
-      "eval_LOCATION_recall": 0.8356164383561644,
-      "eval_ORGANIZATION_f1": 0.7194244604316546,
+      "eval_LOCATION_precision": 0.8227848101265823,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7246376811594203,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6756756756756757,
+      "eval_ORGANIZATION_precision": 0.684931506849315,
       "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8673139158576051,
+      "eval_PERSON_f1": 0.8810289389067525,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8427672955974843,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8509316770186336,
+      "eval_PERSON_recall": 0.9133333333333333,
+      "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.16767063736915588,
-      "eval_overall_accuracy": 0.9618508491262614,
-      "eval_overall_f1": 0.8209366391184574,
-      "eval_overall_precision": 0.7946666666666666,
-      "eval_overall_recall": 0.8490028490028491,
-      "eval_runtime": 1.0071,
-      "eval_samples_per_second": 185.686,
-      "eval_steps_per_second": 2.979,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.1736210733652115,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8365122615803815,
+      "eval_overall_precision": 0.8015665796344648,
+      "eval_overall_recall": 0.8746438746438746,
+      "eval_runtime": 0.3877,
+      "eval_samples_per_second": 482.314,
+      "eval_steps_per_second": 7.738,
       "step": 7632
     },
     {
       "epoch": 73.0,
-      "grad_norm": 0.1150897815823555,
+      "grad_norm": 0.2804931700229645,
       "learning_rate": 1.3500000000000001e-05,
-      "loss": 0.0321,
+      "loss": 0.029,
       "step": 7738
     },
     {
       "epoch": 73.0,
-      "eval_LOCATION_f1": 0.8513513513513513,
+      "eval_LOCATION_f1": 0.8590604026845637,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.84,
-      "eval_LOCATION_recall": 0.863013698630137,
-      "eval_ORGANIZATION_f1": 0.7194244604316546,
+      "eval_LOCATION_precision": 0.8421052631578947,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.76056338028169,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6756756756756757,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.7012987012987013,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_precision": 0.7575757575757576,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17036782205104828,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.8358620689655173,
-      "eval_overall_precision": 0.8101604278074866,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 1.0038,
-      "eval_samples_per_second": 186.293,
-      "eval_steps_per_second": 2.989,
+      "eval_loss": 0.1748729646205902,
+      "eval_overall_accuracy": 0.9648043317745508,
+      "eval_overall_f1": 0.8501362397820162,
+      "eval_overall_precision": 0.814621409921671,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3856,
+      "eval_samples_per_second": 484.921,
+      "eval_steps_per_second": 7.779,
       "step": 7738
     },
     {
       "epoch": 74.0,
-      "grad_norm": 0.2557017505168915,
+      "grad_norm": 0.4218845069408417,
       "learning_rate": 1.3000000000000001e-05,
-      "loss": 0.0306,
+      "loss": 0.0271,
       "step": 7844
     },
     {
       "epoch": 74.0,
-      "eval_LOCATION_f1": 0.8435374149659863,
+      "eval_LOCATION_f1": 0.8724832214765101,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8378378378378378,
-      "eval_LOCATION_recall": 0.8493150684931506,
-      "eval_ORGANIZATION_f1": 0.7101449275362318,
+      "eval_LOCATION_precision": 0.8552631578947368,
+      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6712328767123288,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8745980707395499,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8996763754045306,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.84472049689441,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7812500000000001,
+      "eval_PERSON_precision": 0.8742138364779874,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_precision": 0.78125,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1716192364692688,
-      "eval_overall_accuracy": 0.9628353433423579,
-      "eval_overall_f1": 0.8273972602739725,
-      "eval_overall_precision": 0.7968337730870713,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9992,
-      "eval_samples_per_second": 187.155,
-      "eval_steps_per_second": 3.002,
+      "eval_loss": 0.17355231940746307,
+      "eval_overall_accuracy": 0.9667733202067438,
+      "eval_overall_f1": 0.8532235939643347,
+      "eval_overall_precision": 0.8227513227513228,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3872,
+      "eval_samples_per_second": 482.9,
+      "eval_steps_per_second": 7.747,
       "step": 7844
     },
     {
       "epoch": 75.0,
-      "grad_norm": 0.24300232529640198,
+      "grad_norm": 0.338538259267807,
       "learning_rate": 1.25e-05,
-      "loss": 0.0299,
+      "loss": 0.0283,
       "step": 7950
     },
     {
       "epoch": 75.0,
-      "eval_LOCATION_f1": 0.8387096774193549,
+      "eval_LOCATION_f1": 0.8684210526315789,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7926829268292683,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7338129496402878,
+      "eval_LOCATION_precision": 0.8354430379746836,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7464788732394366,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6891891891891891,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.6883116883116883,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.8387096774193549,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7878787878787878,
-      "eval_QUANTITY_recall": 0.896551724137931,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8450704225352113,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.8108108108108109,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17053084075450897,
-      "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8392370572207084,
-      "eval_overall_precision": 0.804177545691906,
-      "eval_overall_recall": 0.8774928774928775,
-      "eval_runtime": 0.9826,
-      "eval_samples_per_second": 190.304,
-      "eval_steps_per_second": 3.053,
+      "eval_loss": 0.17770251631736755,
+      "eval_overall_accuracy": 0.9643120846665025,
+      "eval_overall_f1": 0.8532608695652173,
+      "eval_overall_precision": 0.8155844155844156,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3855,
+      "eval_samples_per_second": 485.067,
+      "eval_steps_per_second": 7.782,
       "step": 7950
     },
     {
       "epoch": 76.0,
-      "grad_norm": 0.3287041485309601,
+      "grad_norm": 0.8112855553627014,
       "learning_rate": 1.2e-05,
-      "loss": 0.0283,
+      "loss": 0.0292,
       "step": 8056
     },
     {
       "epoch": 76.0,
-      "eval_LOCATION_f1": 0.8387096774193549,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7926829268292683,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7234042553191491,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7534246575342466,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6710526315789473,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8831168831168831,
+      "eval_ORGANIZATION_precision": 0.6790123456790124,
+      "eval_ORGANIZATION_recall": 0.8461538461538461,
+      "eval_PERSON_f1": 0.9096774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8607594936708861,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.88125,
+      "eval_PERSON_recall": 0.94,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8571428571428571,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1747513860464096,
-      "eval_overall_accuracy": 0.9620969726802855,
-      "eval_overall_f1": 0.8315217391304348,
-      "eval_overall_precision": 0.7948051948051948,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9912,
-      "eval_samples_per_second": 188.663,
-      "eval_steps_per_second": 3.027,
+      "eval_loss": 0.1849164366722107,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8544474393530996,
+      "eval_overall_precision": 0.8107416879795396,
+      "eval_overall_recall": 0.9031339031339032,
+      "eval_runtime": 0.3869,
+      "eval_samples_per_second": 483.365,
+      "eval_steps_per_second": 7.755,
       "step": 8056
     },
     {
       "epoch": 77.0,
-      "grad_norm": 1.2225937843322754,
+      "grad_norm": 0.6936742067337036,
       "learning_rate": 1.1500000000000002e-05,
-      "loss": 0.0303,
+      "loss": 0.0287,
       "step": 8162
     },
     {
       "epoch": 77.0,
-      "eval_LOCATION_f1": 0.8333333333333334,
+      "eval_LOCATION_f1": 0.8724832214765101,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7831325301204819,
+      "eval_LOCATION_precision": 0.8552631578947368,
       "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.72992700729927,
+      "eval_ORGANIZATION_f1": 0.7552447552447553,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6944444444444444,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.6923076923076923,
+      "eval_ORGANIZATION_recall": 0.8307692307692308,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17066791653633118,
-      "eval_overall_accuracy": 0.963081466896382,
-      "eval_overall_f1": 0.8333333333333334,
-      "eval_overall_precision": 0.800524934383202,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9877,
-      "eval_samples_per_second": 189.325,
-      "eval_steps_per_second": 3.037,
+      "eval_loss": 0.17852583527565002,
+      "eval_overall_accuracy": 0.9640659611124784,
+      "eval_overall_f1": 0.8517006802721088,
+      "eval_overall_precision": 0.8151041666666666,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3948,
+      "eval_samples_per_second": 473.674,
+      "eval_steps_per_second": 7.599,
       "step": 8162
     },
     {
       "epoch": 78.0,
-      "grad_norm": 1.3090835809707642,
+      "grad_norm": 0.9536997675895691,
       "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.0279,
+      "loss": 0.0258,
       "step": 8268
     },
     {
       "epoch": 78.0,
-      "eval_LOCATION_f1": 0.8387096774193549,
+      "eval_LOCATION_f1": 0.870748299319728,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7926829268292683,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7205882352941175,
+      "eval_LOCATION_precision": 0.8648648648648649,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7647058823529411,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6901408450704225,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8896103896103896,
+      "eval_ORGANIZATION_precision": 0.7323943661971831,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.8925081433224756,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8670886075949367,
+      "eval_PERSON_precision": 0.8726114649681529,
       "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.8064516129032258,
+      "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_precision": 0.78125,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8571428571428571,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17071382701396942,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8356164383561645,
-      "eval_overall_precision": 0.8047493403693932,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 0.9773,
-      "eval_samples_per_second": 191.338,
-      "eval_steps_per_second": 3.07,
+      "eval_TIME_precision": 0.8333333333333334,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.17140330374240875,
+      "eval_overall_accuracy": 0.9662810730986956,
+      "eval_overall_f1": 0.854368932038835,
+      "eval_overall_precision": 0.8324324324324325,
+      "eval_overall_recall": 0.8774928774928775,
+      "eval_runtime": 0.3858,
+      "eval_samples_per_second": 484.686,
+      "eval_steps_per_second": 7.776,
       "step": 8268
     },
     {
       "epoch": 79.0,
-      "grad_norm": 0.48243507742881775,
+      "grad_norm": 0.6820342540740967,
       "learning_rate": 1.05e-05,
-      "loss": 0.0291,
+      "loss": 0.0275,
       "step": 8374
     },
     {
       "epoch": 79.0,
-      "eval_LOCATION_f1": 0.8496732026143791,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8125,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7313432835820897,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7101449275362319,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8802588996763754,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8553459119496856,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17151491343975067,
-      "eval_overall_accuracy": 0.9643120846665025,
-      "eval_overall_f1": 0.8324175824175825,
-      "eval_overall_precision": 0.8037135278514589,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9871,
-      "eval_samples_per_second": 189.447,
-      "eval_steps_per_second": 3.039,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.17593428492546082,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8401084010840107,
+      "eval_overall_precision": 0.8010335917312662,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3867,
+      "eval_samples_per_second": 483.59,
+      "eval_steps_per_second": 7.758,
       "step": 8374
     },
     {
       "epoch": 80.0,
-      "grad_norm": 0.3613625466823578,
+      "grad_norm": 0.6609770655632019,
       "learning_rate": 1e-05,
-      "loss": 0.0291,
+      "loss": 0.0288,
       "step": 8480
     },
     {
       "epoch": 80.0,
-      "eval_LOCATION_f1": 0.846153846153846,
+      "eval_LOCATION_f1": 0.8859060402684564,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7951807228915663,
+      "eval_LOCATION_precision": 0.868421052631579,
       "eval_LOCATION_recall": 0.9041095890410958,
-      "eval_ORGANIZATION_f1": 0.7194244604316546,
+      "eval_ORGANIZATION_f1": 0.7285714285714285,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6756756756756757,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8823529411764707,
+      "eval_ORGANIZATION_precision": 0.68,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9025974025974027,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8653846153846154,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.879746835443038,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8524590163934426,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.8125,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7894736842105263,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.17464391887187958,
-      "eval_overall_accuracy": 0.963081466896382,
-      "eval_overall_f1": 0.8344733242134063,
-      "eval_overall_precision": 0.8026315789473685,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 1.0025,
-      "eval_samples_per_second": 186.532,
-      "eval_steps_per_second": 2.992,
+      "eval_loss": 0.17412517964839935,
+      "eval_overall_accuracy": 0.9640659611124784,
+      "eval_overall_f1": 0.854794520547945,
+      "eval_overall_precision": 0.8232189973614775,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3871,
+      "eval_samples_per_second": 483.069,
+      "eval_steps_per_second": 7.75,
       "step": 8480
     },
     {
       "epoch": 81.0,
-      "grad_norm": 0.47663748264312744,
+      "grad_norm": 0.8274201154708862,
       "learning_rate": 9.5e-06,
-      "loss": 0.0264,
+      "loss": 0.0283,
       "step": 8586
     },
     {
       "epoch": 81.0,
-      "eval_LOCATION_f1": 0.8333333333333334,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7831325301204819,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7205882352941175,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7536231884057971,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6901408450704225,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8823529411764707,
+      "eval_ORGANIZATION_precision": 0.7123287671232876,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8653846153846154,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8571428571428571,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17564432322978973,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.17435409128665924,
       "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8285322359396434,
-      "eval_overall_precision": 0.798941798941799,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9953,
-      "eval_samples_per_second": 187.89,
-      "eval_steps_per_second": 3.014,
+      "eval_overall_f1": 0.8451086956521738,
+      "eval_overall_precision": 0.8077922077922078,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3874,
+      "eval_samples_per_second": 482.691,
+      "eval_steps_per_second": 7.744,
       "step": 8586
     },
     {
       "epoch": 82.0,
-      "grad_norm": 0.739096462726593,
+      "grad_norm": 1.3721864223480225,
       "learning_rate": 9e-06,
-      "loss": 0.0266,
+      "loss": 0.027,
       "step": 8692
     },
     {
       "epoch": 82.0,
-      "eval_LOCATION_f1": 0.8333333333333334,
+      "eval_LOCATION_f1": 0.855263157894737,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7831325301204819,
+      "eval_LOCATION_precision": 0.8227848101265823,
       "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.725925925925926,
+      "eval_ORGANIZATION_f1": 0.7285714285714285,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8823529411764707,
+      "eval_ORGANIZATION_precision": 0.68,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8990228013029316,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8653846153846154,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8789808917197452,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.861111111111111,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17531628906726837,
+      "eval_TIME_precision": 0.8157894736842105,
+      "eval_TIME_recall": 0.9117647058823529,
+      "eval_loss": 0.17581164836883545,
       "eval_overall_accuracy": 0.9643120846665025,
-      "eval_overall_f1": 0.8308115543328748,
-      "eval_overall_precision": 0.8031914893617021,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9906,
-      "eval_samples_per_second": 188.775,
-      "eval_steps_per_second": 3.028,
+      "eval_overall_f1": 0.8485675306957708,
+      "eval_overall_precision": 0.8141361256544503,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3861,
+      "eval_samples_per_second": 484.302,
+      "eval_steps_per_second": 7.77,
       "step": 8692
     },
     {
       "epoch": 83.0,
-      "grad_norm": 3.826598644256592,
+      "grad_norm": 1.29371178150177,
       "learning_rate": 8.500000000000002e-06,
-      "loss": 0.0248,
+      "loss": 0.0254,
       "step": 8798
     },
     {
       "epoch": 83.0,
-      "eval_LOCATION_f1": 0.8441558441558442,
+      "eval_LOCATION_f1": 0.8533333333333334,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8024691358024691,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7480916030534351,
+      "eval_LOCATION_precision": 0.8311688311688312,
+      "eval_LOCATION_recall": 0.8767123287671232,
+      "eval_ORGANIZATION_f1": 0.7222222222222222,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7424242424242424,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8729641693811074,
+      "eval_ORGANIZATION_precision": 0.6582278481012658,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.896103896103896,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8535031847133758,
-      "eval_PERSON_recall": 0.8933333333333333,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8734177215189873,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17535383999347687,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.8326417704011064,
-      "eval_overall_precision": 0.8091397849462365,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 1.0064,
-      "eval_samples_per_second": 185.819,
-      "eval_steps_per_second": 2.981,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18245863914489746,
+      "eval_overall_accuracy": 0.9628353433423579,
+      "eval_overall_f1": 0.8373983739837397,
+      "eval_overall_precision": 0.7984496124031008,
+      "eval_overall_recall": 0.8803418803418803,
+      "eval_runtime": 0.3873,
+      "eval_samples_per_second": 482.882,
+      "eval_steps_per_second": 7.747,
       "step": 8798
     },
     {
       "epoch": 84.0,
-      "grad_norm": 0.8354291915893555,
+      "grad_norm": 0.759432852268219,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.027,
+      "loss": 0.0267,
       "step": 8904
     },
     {
       "epoch": 84.0,
-      "eval_LOCATION_f1": 0.8609271523178808,
+      "eval_LOCATION_f1": 0.8684210526315789,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8333333333333334,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7246376811594203,
+      "eval_LOCATION_precision": 0.8354430379746836,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7412587412587412,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.684931506849315,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8831168831168831,
+      "eval_ORGANIZATION_precision": 0.6794871794871795,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.887459807073955,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8607594936708861,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8571428571428571,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8695652173913043,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8219178082191781,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8571428571428571,
+      "eval_TIME_precision": 0.7692307692307693,
       "eval_TIME_recall": 0.8823529411764706,
-      "eval_loss": 0.1745665818452835,
-      "eval_overall_accuracy": 0.9648043317745508,
-      "eval_overall_f1": 0.8390646492434664,
-      "eval_overall_precision": 0.8111702127659575,
-      "eval_overall_recall": 0.8689458689458689,
-      "eval_runtime": 1.0704,
-      "eval_samples_per_second": 174.7,
-      "eval_steps_per_second": 2.803,
+      "eval_loss": 0.18526050448417664,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8398384925975775,
+      "eval_overall_precision": 0.7959183673469388,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3876,
+      "eval_samples_per_second": 482.481,
+      "eval_steps_per_second": 7.74,
       "step": 8904
     },
     {
       "epoch": 85.0,
-      "grad_norm": 1.275327444076538,
+      "grad_norm": 0.7862147092819214,
       "learning_rate": 7.5e-06,
-      "loss": 0.0278,
+      "loss": 0.0244,
       "step": 9010
     },
     {
       "epoch": 85.0,
-      "eval_LOCATION_f1": 0.8387096774193549,
+      "eval_LOCATION_f1": 0.8627450980392157,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7926829268292683,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7153284671532848,
+      "eval_LOCATION_precision": 0.825,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7142857142857142,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6805555555555556,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8810289389067525,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8509316770186336,
-      "eval_PERSON_recall": 0.9133333333333333,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.18171997368335724,
-      "eval_overall_accuracy": 0.9628353433423579,
-      "eval_overall_f1": 0.8283378746594006,
-      "eval_overall_precision": 0.793733681462141,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.988,
-      "eval_samples_per_second": 189.269,
-      "eval_steps_per_second": 3.036,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18103930354118347,
+      "eval_overall_accuracy": 0.9643120846665025,
+      "eval_overall_f1": 0.8412483039348712,
+      "eval_overall_precision": 0.8031088082901554,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.392,
+      "eval_samples_per_second": 477.052,
+      "eval_steps_per_second": 7.653,
       "step": 9010
     },
     {
       "epoch": 86.0,
-      "grad_norm": 0.6492407321929932,
+      "grad_norm": 0.45432260632514954,
       "learning_rate": 7.000000000000001e-06,
-      "loss": 0.0266,
+      "loss": 0.0241,
       "step": 9116
     },
     {
       "epoch": 86.0,
-      "eval_LOCATION_f1": 0.8387096774193549,
+      "eval_LOCATION_f1": 0.881578947368421,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7926829268292683,
-      "eval_LOCATION_recall": 0.8904109589041096,
+      "eval_LOCATION_precision": 0.8481012658227848,
+      "eval_LOCATION_recall": 0.9178082191780822,
       "eval_ORGANIZATION_f1": 0.7313432835820897,
       "eval_ORGANIZATION_number": 65,
       "eval_ORGANIZATION_precision": 0.7101449275362319,
       "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_PERSON_f1": 0.8932038834951457,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.8679245283018868,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8253968253968255,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
-      "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7647058823529411,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1775309443473816,
-      "eval_overall_accuracy": 0.9648043317745508,
-      "eval_overall_f1": 0.8324175824175825,
-      "eval_overall_precision": 0.8037135278514589,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9952,
-      "eval_samples_per_second": 187.906,
-      "eval_steps_per_second": 3.015,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.17840129137039185,
+      "eval_overall_accuracy": 0.965296578882599,
+      "eval_overall_f1": 0.8493150684931507,
+      "eval_overall_precision": 0.8179419525065963,
+      "eval_overall_recall": 0.8831908831908832,
+      "eval_runtime": 0.3875,
+      "eval_samples_per_second": 482.525,
+      "eval_steps_per_second": 7.741,
       "step": 9116
     },
     {
       "epoch": 87.0,
-      "grad_norm": 0.48030099272727966,
+      "grad_norm": 1.0281379222869873,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.0261,
+      "loss": 0.0257,
       "step": 9222
     },
     {
       "epoch": 87.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8758169934640524,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7183098591549296,
+      "eval_LOCATION_precision": 0.8375,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7272727272727272,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6623376623376623,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6666666666666666,
+      "eval_ORGANIZATION_recall": 0.8,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17771387100219727,
-      "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8340192043895748,
-      "eval_overall_precision": 0.8042328042328042,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9953,
-      "eval_samples_per_second": 187.876,
-      "eval_steps_per_second": 3.014,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.1823926866054535,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.844804318488529,
+      "eval_overall_precision": 0.8025641025641026,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3876,
+      "eval_samples_per_second": 482.435,
+      "eval_steps_per_second": 7.74,
       "step": 9222
     },
     {
       "epoch": 88.0,
-      "grad_norm": 0.7421311736106873,
+      "grad_norm": 0.4464998245239258,
       "learning_rate": 6e-06,
-      "loss": 0.0276,
+      "loss": 0.0253,
       "step": 9328
     },
     {
       "epoch": 88.0,
-      "eval_LOCATION_f1": 0.847682119205298,
+      "eval_LOCATION_f1": 0.8874172185430463,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8205128205128205,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.72992700729927,
+      "eval_LOCATION_precision": 0.8589743589743589,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6944444444444444,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.7812500000000001,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7142857142857143,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1769006848335266,
-      "eval_overall_accuracy": 0.9633275904504061,
-      "eval_overall_f1": 0.8319559228650139,
-      "eval_overall_precision": 0.8053333333333333,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.998,
-      "eval_samples_per_second": 187.379,
-      "eval_steps_per_second": 3.006,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18025493621826172,
+      "eval_overall_accuracy": 0.9640659611124784,
+      "eval_overall_f1": 0.8443843031123139,
+      "eval_overall_precision": 0.8041237113402062,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3861,
+      "eval_samples_per_second": 484.269,
+      "eval_steps_per_second": 7.769,
       "step": 9328
     },
     {
       "epoch": 89.0,
-      "grad_norm": 0.4191894829273224,
+      "grad_norm": 1.2788630723953247,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.0261,
+      "loss": 0.026,
       "step": 9434
     },
     {
       "epoch": 89.0,
-      "eval_LOCATION_f1": 0.8258064516129032,
+      "eval_LOCATION_f1": 0.8758169934640524,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7804878048780488,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7153284671532848,
+      "eval_LOCATION_precision": 0.8375,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6805555555555556,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.75,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
+      "eval_QUANTITY_precision": 0.6857142857142857,
       "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1796666532754898,
-      "eval_overall_accuracy": 0.963081466896382,
-      "eval_overall_f1": 0.8257887517146777,
-      "eval_overall_precision": 0.7962962962962963,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 0.9969,
-      "eval_samples_per_second": 187.575,
-      "eval_steps_per_second": 3.009,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18378011882305145,
+      "eval_overall_accuracy": 0.9623430962343096,
+      "eval_overall_f1": 0.8394062078272606,
+      "eval_overall_precision": 0.7974358974358975,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3874,
+      "eval_samples_per_second": 482.728,
+      "eval_steps_per_second": 7.744,
       "step": 9434
     },
     {
       "epoch": 90.0,
-      "grad_norm": 0.639091968536377,
+      "grad_norm": 0.9298455715179443,
       "learning_rate": 5e-06,
-      "loss": 0.0251,
+      "loss": 0.0246,
       "step": 9540
     },
     {
       "epoch": 90.0,
-      "eval_LOCATION_f1": 0.8496732026143791,
+      "eval_LOCATION_f1": 0.8933333333333333,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8125,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8701298701298701,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.8903225806451613,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8625,
+      "eval_PERSON_recall": 0.92,
+      "eval_QUANTITY_f1": 0.8387096774193549,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7878787878787878,
+      "eval_QUANTITY_recall": 0.896551724137931,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17721274495124817,
-      "eval_overall_accuracy": 0.9650504553285749,
-      "eval_overall_f1": 0.837465564738292,
-      "eval_overall_precision": 0.8106666666666666,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 0.9991,
-      "eval_samples_per_second": 187.173,
-      "eval_steps_per_second": 3.003,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18070749938488007,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8478260869565218,
+      "eval_overall_precision": 0.8103896103896104,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3896,
+      "eval_samples_per_second": 480.006,
+      "eval_steps_per_second": 7.701,
       "step": 9540
     },
     {
       "epoch": 91.0,
-      "grad_norm": 0.6990874409675598,
+      "grad_norm": 0.9396980404853821,
       "learning_rate": 4.5e-06,
-      "loss": 0.0243,
+      "loss": 0.023,
       "step": 9646
     },
     {
       "epoch": 91.0,
-      "eval_LOCATION_f1": 0.8311688311688312,
+      "eval_LOCATION_f1": 0.8874172185430463,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.7901234567901234,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7368421052631577,
+      "eval_LOCATION_precision": 0.8589743589743589,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7042253521126761,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7205882352941176,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.6493506493506493,
+      "eval_ORGANIZATION_recall": 0.7692307692307693,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7741935483870968,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.7272727272727273,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17922192811965942,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.830344827586207,
-      "eval_overall_precision": 0.8048128342245989,
-      "eval_overall_recall": 0.8575498575498576,
-      "eval_runtime": 0.9978,
-      "eval_samples_per_second": 187.404,
-      "eval_steps_per_second": 3.006,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18405993282794952,
+      "eval_overall_accuracy": 0.9640659611124784,
+      "eval_overall_f1": 0.8466757123473541,
+      "eval_overall_precision": 0.8082901554404145,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.39,
+      "eval_samples_per_second": 479.51,
+      "eval_steps_per_second": 7.693,
       "step": 9646
     },
     {
       "epoch": 92.0,
-      "grad_norm": 0.31417253613471985,
+      "grad_norm": 0.77614825963974,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.0246,
+      "loss": 0.0235,
       "step": 9752
     },
     {
       "epoch": 92.0,
-      "eval_LOCATION_f1": 0.8496732026143791,
+      "eval_LOCATION_f1": 0.881578947368421,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8125,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.725925925925926,
+      "eval_LOCATION_precision": 0.8481012658227848,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7285714285714285,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.7,
-      "eval_ORGANIZATION_recall": 0.7538461538461538,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.68,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7352941176470589,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17927689850330353,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.833103448275862,
-      "eval_overall_precision": 0.8074866310160428,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 1.0031,
-      "eval_samples_per_second": 186.424,
-      "eval_steps_per_second": 2.991,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18313735723495483,
+      "eval_overall_accuracy": 0.9648043317745508,
+      "eval_overall_f1": 0.8505434782608695,
+      "eval_overall_precision": 0.812987012987013,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3921,
+      "eval_samples_per_second": 476.865,
+      "eval_steps_per_second": 7.65,
       "step": 9752
     },
     {
       "epoch": 93.0,
-      "grad_norm": 0.2049255222082138,
+      "grad_norm": 0.6140993237495422,
       "learning_rate": 3.5000000000000004e-06,
-      "loss": 0.0252,
+      "loss": 0.0228,
       "step": 9858
     },
     {
       "epoch": 93.0,
-      "eval_LOCATION_f1": 0.8609271523178808,
+      "eval_LOCATION_f1": 0.8627450980392157,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8333333333333334,
-      "eval_LOCATION_recall": 0.8904109589041096,
-      "eval_ORGANIZATION_f1": 0.7338129496402878,
+      "eval_LOCATION_precision": 0.825,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7361111111111112,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6891891891891891,
-      "eval_ORGANIZATION_recall": 0.7846153846153846,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6708860759493671,
+      "eval_ORGANIZATION_recall": 0.8153846153846154,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
       "eval_QUANTITY_f1": 0.8196721311475409,
       "eval_QUANTITY_number": 29,
       "eval_QUANTITY_precision": 0.78125,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1780104637145996,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18428488075733185,
       "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8418156808803302,
-      "eval_overall_precision": 0.8138297872340425,
-      "eval_overall_recall": 0.8717948717948718,
-      "eval_runtime": 0.9912,
-      "eval_samples_per_second": 188.665,
-      "eval_steps_per_second": 3.027,
+      "eval_overall_f1": 0.8497970230040596,
+      "eval_overall_precision": 0.8092783505154639,
+      "eval_overall_recall": 0.8945868945868946,
+      "eval_runtime": 0.3933,
+      "eval_samples_per_second": 475.405,
+      "eval_steps_per_second": 7.627,
       "step": 9858
     },
     {
       "epoch": 94.0,
-      "grad_norm": 0.3145851790904999,
+      "grad_norm": 0.7894928455352783,
       "learning_rate": 3e-06,
-      "loss": 0.0252,
+      "loss": 0.024,
       "step": 9964
     },
     {
       "epoch": 94.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8758169934640524,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8375,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8794788273615636,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8598726114649682,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.7936507936507937,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_precision": 0.7352941176470589,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17691487073898315,
-      "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8381742738589213,
-      "eval_overall_precision": 0.8145161290322581,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9959,
-      "eval_samples_per_second": 187.776,
-      "eval_steps_per_second": 3.012,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18454931676387787,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8470906630581868,
+      "eval_overall_precision": 0.8067010309278351,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.394,
+      "eval_samples_per_second": 474.646,
+      "eval_steps_per_second": 7.615,
       "step": 9964
     },
     {
       "epoch": 95.0,
-      "grad_norm": 0.3581851124763489,
+      "grad_norm": 0.5310998558998108,
       "learning_rate": 2.5e-06,
-      "loss": 0.0241,
+      "loss": 0.024,
       "step": 10070
     },
     {
       "epoch": 95.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8874172185430463,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.72992700729927,
+      "eval_LOCATION_precision": 0.8589743589743589,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7132867132867133,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.6944444444444444,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6538461538461539,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.9061488673139159,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.8196721311475409,
+      "eval_PERSON_precision": 0.8805031446540881,
+      "eval_PERSON_recall": 0.9333333333333333,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.78125,
+      "eval_QUANTITY_precision": 0.7575757575757576,
       "eval_QUANTITY_recall": 0.8620689655172413,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17827032506465912,
-      "eval_overall_accuracy": 0.9635737140044303,
-      "eval_overall_f1": 0.8397790055248618,
-      "eval_overall_precision": 0.8150134048257373,
-      "eval_overall_recall": 0.8660968660968661,
-      "eval_runtime": 1.0048,
-      "eval_samples_per_second": 186.108,
-      "eval_steps_per_second": 2.986,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18248838186264038,
+      "eval_overall_accuracy": 0.9640659611124784,
+      "eval_overall_f1": 0.8493894165535957,
+      "eval_overall_precision": 0.810880829015544,
+      "eval_overall_recall": 0.8917378917378918,
+      "eval_runtime": 0.3923,
+      "eval_samples_per_second": 476.645,
+      "eval_steps_per_second": 7.647,
       "step": 10070
     },
     {
       "epoch": 96.0,
-      "grad_norm": 0.5018787384033203,
+      "grad_norm": 0.7802311778068542,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.0239,
+      "loss": 0.0247,
       "step": 10176
     },
     {
       "epoch": 96.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8741721854304636,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8461538461538461,
+      "eval_LOCATION_recall": 0.9041095890410958,
+      "eval_ORGANIZATION_f1": 0.7083333333333334,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8766233766233766,
+      "eval_ORGANIZATION_precision": 0.6455696202531646,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8544303797468354,
-      "eval_PERSON_recall": 0.9,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1762346476316452,
-      "eval_overall_accuracy": 0.9638198375584544,
-      "eval_overall_f1": 0.8342541436464089,
-      "eval_overall_precision": 0.8096514745308311,
-      "eval_overall_recall": 0.8603988603988604,
-      "eval_runtime": 0.9962,
-      "eval_samples_per_second": 187.711,
-      "eval_steps_per_second": 3.011,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18491245806217194,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8416779431664412,
+      "eval_overall_precision": 0.8015463917525774,
+      "eval_overall_recall": 0.886039886039886,
+      "eval_runtime": 0.3899,
+      "eval_samples_per_second": 479.554,
+      "eval_steps_per_second": 7.693,
       "step": 10176
     },
     {
       "epoch": 97.0,
-      "grad_norm": 0.7310182452201843,
+      "grad_norm": 0.7217825055122375,
       "learning_rate": 1.5e-06,
-      "loss": 0.0226,
+      "loss": 0.0236,
       "step": 10282
     },
     {
       "epoch": 97.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.881578947368421,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8481012658227848,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7234042553191491,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6710526315789473,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17845137417316437,
-      "eval_overall_accuracy": 0.9643120846665025,
-      "eval_overall_f1": 0.8381742738589213,
-      "eval_overall_precision": 0.8145161290322581,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 1.0543,
-      "eval_samples_per_second": 177.368,
-      "eval_steps_per_second": 2.845,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18391452729701996,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8466757123473541,
+      "eval_overall_precision": 0.8082901554404145,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3979,
+      "eval_samples_per_second": 469.952,
+      "eval_steps_per_second": 7.539,
       "step": 10282
     },
     {
       "epoch": 98.0,
-      "grad_norm": 0.48724424839019775,
+      "grad_norm": 0.31892678141593933,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 0.0244,
+      "loss": 0.0228,
       "step": 10388
     },
     {
       "epoch": 98.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8758169934640524,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8375,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7183098591549296,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6623376623376623,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8285714285714286,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8055555555555556,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17859233915805817,
-      "eval_overall_accuracy": 0.9640659611124784,
-      "eval_overall_f1": 0.8370165745856354,
-      "eval_overall_precision": 0.8123324396782842,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9905,
-      "eval_samples_per_second": 188.793,
-      "eval_steps_per_second": 3.029,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18566611409187317,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8443843031123139,
+      "eval_overall_precision": 0.8041237113402062,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3908,
+      "eval_samples_per_second": 478.561,
+      "eval_steps_per_second": 7.677,
       "step": 10388
     },
     {
       "epoch": 99.0,
-      "grad_norm": 0.3531472086906433,
+      "grad_norm": 1.421492099761963,
       "learning_rate": 5.000000000000001e-07,
-      "loss": 0.0244,
+      "loss": 0.025,
       "step": 10494
     },
     {
       "epoch": 99.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.881578947368421,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8481012658227848,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7132867132867133,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6538461538461539,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.17733581364154816,
-      "eval_overall_accuracy": 0.9645582082205267,
-      "eval_overall_f1": 0.8381742738589213,
-      "eval_overall_precision": 0.8145161290322581,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.992,
-      "eval_samples_per_second": 188.514,
-      "eval_steps_per_second": 3.024,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.18506501615047455,
+      "eval_overall_accuracy": 0.9638198375584544,
+      "eval_overall_f1": 0.8443843031123139,
+      "eval_overall_precision": 0.8041237113402062,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3928,
+      "eval_samples_per_second": 476.04,
+      "eval_steps_per_second": 7.637,
       "step": 10494
     },
     {
       "epoch": 100.0,
-      "grad_norm": 0.8635873794555664,
+      "grad_norm": 0.21340006589889526,
       "learning_rate": 0.0,
-      "loss": 0.0265,
+      "loss": 0.023,
       "step": 10600
     },
     {
       "epoch": 100.0,
-      "eval_LOCATION_f1": 0.8533333333333334,
+      "eval_LOCATION_f1": 0.8758169934640524,
       "eval_LOCATION_number": 73,
-      "eval_LOCATION_precision": 0.8311688311688312,
-      "eval_LOCATION_recall": 0.8767123287671232,
-      "eval_ORGANIZATION_f1": 0.7352941176470589,
+      "eval_LOCATION_precision": 0.8375,
+      "eval_LOCATION_recall": 0.9178082191780822,
+      "eval_ORGANIZATION_f1": 0.7132867132867133,
       "eval_ORGANIZATION_number": 65,
-      "eval_ORGANIZATION_precision": 0.704225352112676,
-      "eval_ORGANIZATION_recall": 0.7692307692307693,
-      "eval_PERSON_f1": 0.8859934853420196,
+      "eval_ORGANIZATION_precision": 0.6538461538461539,
+      "eval_ORGANIZATION_recall": 0.7846153846153846,
+      "eval_PERSON_f1": 0.896774193548387,
       "eval_PERSON_number": 150,
-      "eval_PERSON_precision": 0.8662420382165605,
-      "eval_PERSON_recall": 0.9066666666666666,
-      "eval_QUANTITY_f1": 0.7868852459016394,
+      "eval_PERSON_precision": 0.86875,
+      "eval_PERSON_recall": 0.9266666666666666,
+      "eval_QUANTITY_f1": 0.8064516129032258,
       "eval_QUANTITY_number": 29,
-      "eval_QUANTITY_precision": 0.75,
-      "eval_QUANTITY_recall": 0.8275862068965517,
-      "eval_TIME_f1": 0.8405797101449276,
+      "eval_QUANTITY_precision": 0.7575757575757576,
+      "eval_QUANTITY_recall": 0.8620689655172413,
+      "eval_TIME_f1": 0.8333333333333333,
       "eval_TIME_number": 34,
-      "eval_TIME_precision": 0.8285714285714286,
-      "eval_TIME_recall": 0.8529411764705882,
-      "eval_loss": 0.1773531436920166,
-      "eval_overall_accuracy": 0.9643120846665025,
-      "eval_overall_f1": 0.8381742738589213,
-      "eval_overall_precision": 0.8145161290322581,
-      "eval_overall_recall": 0.8632478632478633,
-      "eval_runtime": 0.9958,
-      "eval_samples_per_second": 187.787,
-      "eval_steps_per_second": 3.013,
+      "eval_TIME_precision": 0.7894736842105263,
+      "eval_TIME_recall": 0.8823529411764706,
+      "eval_loss": 0.1850329041481018,
+      "eval_overall_accuracy": 0.9635737140044303,
+      "eval_overall_f1": 0.8432432432432433,
+      "eval_overall_precision": 0.8020565552699229,
+      "eval_overall_recall": 0.8888888888888888,
+      "eval_runtime": 0.3916,
+      "eval_samples_per_second": 477.569,
+      "eval_steps_per_second": 7.662,
       "step": 10600
     },
     {
       "epoch": 100.0,
       "step": 10600,
       "total_flos": 5043631411410816.0,
-      "train_loss": 0.07336943858074692,
-      "train_runtime": 1365.3768,
-      "train_samples_per_second": 123.629,
-      "train_steps_per_second": 7.763
+      "train_loss": 0.07131878634668747,
+      "train_runtime": 622.8213,
+      "train_samples_per_second": 271.025,
+      "train_steps_per_second": 17.019
     }
   ],
   "logging_steps": 500,