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": 1.569076418876648,
+      "grad_norm": 1.4739856719970703,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 1.1639,
+      "loss": 1.1566,
       "step": 96
     },
     {
@@ -29,21 +29,21 @@
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.0,
       "eval_PERSON_recall": 0.0,
-      "eval_loss": 0.6982586979866028,
+      "eval_loss": 0.695199191570282,
       "eval_overall_accuracy": 0.8373377520022093,
       "eval_overall_f1": 0.0,
       "eval_overall_precision": 0.0,
       "eval_overall_recall": 0.0,
-      "eval_runtime": 0.5919,
-      "eval_samples_per_second": 287.192,
-      "eval_steps_per_second": 5.068,
+      "eval_runtime": 0.2694,
+      "eval_samples_per_second": 630.973,
+      "eval_steps_per_second": 11.135,
       "step": 96
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.9749519228935242,
+      "grad_norm": 0.9800453186035156,
       "learning_rate": 4.9e-05,
-      "loss": 0.6685,
+      "loss": 0.6676,
       "step": 192
     },
     {
@@ -60,21 +60,21 @@
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.0,
       "eval_PERSON_recall": 0.0,
-      "eval_loss": 0.565019428730011,
+      "eval_loss": 0.5652989745140076,
       "eval_overall_accuracy": 0.8376139188069595,
       "eval_overall_f1": 0.0,
       "eval_overall_precision": 0.0,
       "eval_overall_recall": 0.0,
-      "eval_runtime": 0.6878,
-      "eval_samples_per_second": 247.148,
-      "eval_steps_per_second": 4.361,
+      "eval_runtime": 0.2734,
+      "eval_samples_per_second": 621.732,
+      "eval_steps_per_second": 10.972,
       "step": 192
     },
     {
       "epoch": 3.0,
-      "grad_norm": 1.0692148208618164,
+      "grad_norm": 1.1218823194503784,
       "learning_rate": 4.85e-05,
-      "loss": 0.553,
+      "loss": 0.5559,
       "step": 288
     },
     {
@@ -83,525 +83,525 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.0,
       "eval_LOCATION_recall": 0.0,
-      "eval_ORGANIZATION_f1": 0.07407407407407407,
-      "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.3888888888888889,
-      "eval_ORGANIZATION_recall": 0.04093567251461988,
-      "eval_PERSON_f1": 0.10526315789473685,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.225,
-      "eval_PERSON_recall": 0.06870229007633588,
-      "eval_loss": 0.4425106644630432,
-      "eval_overall_accuracy": 0.8470035901684617,
-      "eval_overall_f1": 0.06866952789699571,
-      "eval_overall_precision": 0.26229508196721313,
-      "eval_overall_recall": 0.03950617283950617,
-      "eval_runtime": 0.6213,
-      "eval_samples_per_second": 273.628,
-      "eval_steps_per_second": 4.829,
+      "eval_ORGANIZATION_f1": 0.0641711229946524,
+      "eval_ORGANIZATION_number": 171,
+      "eval_ORGANIZATION_precision": 0.375,
+      "eval_ORGANIZATION_recall": 0.03508771929824561,
+      "eval_PERSON_f1": 0.08588957055214724,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.21875,
+      "eval_PERSON_recall": 0.05343511450381679,
+      "eval_loss": 0.44871625304222107,
+      "eval_overall_accuracy": 0.8456227561447114,
+      "eval_overall_f1": 0.05714285714285715,
+      "eval_overall_precision": 0.26,
+      "eval_overall_recall": 0.03209876543209877,
+      "eval_runtime": 0.2727,
+      "eval_samples_per_second": 623.419,
+      "eval_steps_per_second": 11.002,
       "step": 288
     },
     {
       "epoch": 4.0,
-      "grad_norm": 0.8610777258872986,
+      "grad_norm": 0.906775176525116,
       "learning_rate": 4.8e-05,
-      "loss": 0.4403,
+      "loss": 0.4455,
       "step": 384
     },
     {
       "epoch": 4.0,
-      "eval_LOCATION_f1": 0.078125,
+      "eval_LOCATION_f1": 0.07874015748031495,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.2,
+      "eval_LOCATION_precision": 0.20833333333333334,
       "eval_LOCATION_recall": 0.04854368932038835,
-      "eval_ORGANIZATION_f1": 0.29078014184397166,
-      "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.36936936936936937,
-      "eval_ORGANIZATION_recall": 0.23976608187134502,
-      "eval_PERSON_f1": 0.4290657439446367,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.3924050632911392,
-      "eval_PERSON_recall": 0.4732824427480916,
-      "eval_loss": 0.33291852474212646,
-      "eval_overall_accuracy": 0.8834576083954708,
-      "eval_overall_f1": 0.3090128755364807,
-      "eval_overall_precision": 0.3673469387755102,
-      "eval_overall_recall": 0.26666666666666666,
-      "eval_runtime": 0.6305,
-      "eval_samples_per_second": 269.628,
-      "eval_steps_per_second": 4.758,
+      "eval_ORGANIZATION_f1": 0.2807017543859649,
+      "eval_ORGANIZATION_number": 171,
+      "eval_ORGANIZATION_precision": 0.3508771929824561,
+      "eval_ORGANIZATION_recall": 0.23391812865497075,
+      "eval_PERSON_f1": 0.4098939929328622,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.3815789473684211,
+      "eval_PERSON_recall": 0.44274809160305345,
+      "eval_loss": 0.33893293142318726,
+      "eval_overall_accuracy": 0.8818006075669704,
+      "eval_overall_f1": 0.2964028776978417,
+      "eval_overall_precision": 0.35517241379310344,
+      "eval_overall_recall": 0.254320987654321,
+      "eval_runtime": 0.2771,
+      "eval_samples_per_second": 613.432,
+      "eval_steps_per_second": 10.825,
       "step": 384
     },
     {
       "epoch": 5.0,
-      "grad_norm": 0.5931413769721985,
+      "grad_norm": 0.6574228405952454,
       "learning_rate": 4.75e-05,
-      "loss": 0.3288,
+      "loss": 0.3416,
       "step": 480
     },
     {
       "epoch": 5.0,
-      "eval_LOCATION_f1": 0.3715846994535519,
-      "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.425,
-      "eval_LOCATION_recall": 0.3300970873786408,
-      "eval_ORGANIZATION_f1": 0.5683646112600537,
-      "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.5247524752475248,
-      "eval_ORGANIZATION_recall": 0.6198830409356725,
-      "eval_PERSON_f1": 0.6107382550335572,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.5449101796407185,
-      "eval_PERSON_recall": 0.6946564885496184,
-      "eval_loss": 0.24550960958003998,
-      "eval_overall_accuracy": 0.9262634631317316,
-      "eval_overall_f1": 0.540983606557377,
-      "eval_overall_precision": 0.5144766146993318,
-      "eval_overall_recall": 0.5703703703703704,
-      "eval_runtime": 0.6072,
-      "eval_samples_per_second": 279.961,
-      "eval_steps_per_second": 4.94,
+      "eval_LOCATION_f1": 0.3157894736842105,
+      "eval_LOCATION_number": 103,
+      "eval_LOCATION_precision": 0.39705882352941174,
+      "eval_LOCATION_recall": 0.2621359223300971,
+      "eval_ORGANIZATION_f1": 0.5245901639344263,
+      "eval_ORGANIZATION_number": 171,
+      "eval_ORGANIZATION_precision": 0.49230769230769234,
+      "eval_ORGANIZATION_recall": 0.5614035087719298,
+      "eval_PERSON_f1": 0.5847176079734219,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.5176470588235295,
+      "eval_PERSON_recall": 0.6717557251908397,
+      "eval_loss": 0.25828996300697327,
+      "eval_overall_accuracy": 0.9207401270367301,
+      "eval_overall_f1": 0.503579952267303,
+      "eval_overall_precision": 0.48729792147806006,
+      "eval_overall_recall": 0.5209876543209877,
+      "eval_runtime": 0.2793,
+      "eval_samples_per_second": 608.705,
+      "eval_steps_per_second": 10.742,
       "step": 480
     },
     {
       "epoch": 6.0,
-      "grad_norm": 1.293260097503662,
+      "grad_norm": 1.1292903423309326,
       "learning_rate": 4.7e-05,
-      "loss": 0.2474,
+      "loss": 0.2637,
       "step": 576
     },
     {
       "epoch": 6.0,
-      "eval_LOCATION_f1": 0.626865671641791,
-      "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.6428571428571429,
-      "eval_LOCATION_recall": 0.6116504854368932,
-      "eval_ORGANIZATION_f1": 0.6863270777479892,
-      "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.6336633663366337,
-      "eval_ORGANIZATION_recall": 0.7485380116959064,
-      "eval_PERSON_f1": 0.833922261484099,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.7763157894736842,
-      "eval_PERSON_recall": 0.9007633587786259,
-      "eval_loss": 0.18931446969509125,
-      "eval_overall_accuracy": 0.9505661419497377,
-      "eval_overall_f1": 0.721120186697783,
-      "eval_overall_precision": 0.6836283185840708,
-      "eval_overall_recall": 0.762962962962963,
-      "eval_runtime": 0.5962,
-      "eval_samples_per_second": 285.148,
-      "eval_steps_per_second": 5.032,
+      "eval_LOCATION_f1": 0.6060606060606061,
+      "eval_LOCATION_number": 103,
+      "eval_LOCATION_precision": 0.631578947368421,
+      "eval_LOCATION_recall": 0.5825242718446602,
+      "eval_ORGANIZATION_f1": 0.712401055408971,
+      "eval_ORGANIZATION_number": 171,
+      "eval_ORGANIZATION_precision": 0.6490384615384616,
+      "eval_ORGANIZATION_recall": 0.7894736842105263,
+      "eval_PERSON_f1": 0.767605633802817,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.7124183006535948,
+      "eval_PERSON_recall": 0.8320610687022901,
+      "eval_loss": 0.20057322084903717,
+      "eval_overall_accuracy": 0.9489091411212373,
+      "eval_overall_f1": 0.7061556329849014,
+      "eval_overall_precision": 0.6666666666666666,
+      "eval_overall_recall": 0.7506172839506173,
+      "eval_runtime": 0.2738,
+      "eval_samples_per_second": 620.929,
+      "eval_steps_per_second": 10.958,
       "step": 576
     },
     {
       "epoch": 7.0,
-      "grad_norm": 1.2653672695159912,
+      "grad_norm": 1.1348835229873657,
       "learning_rate": 4.6500000000000005e-05,
-      "loss": 0.1962,
+      "loss": 0.2115,
       "step": 672
     },
     {
       "epoch": 7.0,
-      "eval_LOCATION_f1": 0.7551020408163266,
-      "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.7956989247311828,
-      "eval_LOCATION_recall": 0.7184466019417476,
-      "eval_ORGANIZATION_f1": 0.7526881720430108,
-      "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.6965174129353234,
-      "eval_ORGANIZATION_recall": 0.8187134502923976,
-      "eval_PERSON_f1": 0.9157509157509158,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.8802816901408451,
-      "eval_PERSON_recall": 0.9541984732824428,
-      "eval_loss": 0.1533629298210144,
-      "eval_overall_accuracy": 0.9605081469207402,
-      "eval_overall_f1": 0.8061831153388822,
-      "eval_overall_precision": 0.7775229357798165,
-      "eval_overall_recall": 0.837037037037037,
-      "eval_runtime": 0.5884,
-      "eval_samples_per_second": 288.903,
-      "eval_steps_per_second": 5.098,
+      "eval_LOCATION_f1": 0.712871287128713,
+      "eval_LOCATION_number": 103,
+      "eval_LOCATION_precision": 0.7272727272727273,
+      "eval_LOCATION_recall": 0.6990291262135923,
+      "eval_ORGANIZATION_f1": 0.7540106951871658,
+      "eval_ORGANIZATION_number": 171,
+      "eval_ORGANIZATION_precision": 0.6945812807881774,
+      "eval_ORGANIZATION_recall": 0.8245614035087719,
+      "eval_PERSON_f1": 0.8945454545454545,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.8541666666666666,
+      "eval_PERSON_recall": 0.9389312977099237,
+      "eval_loss": 0.1648775041103363,
+      "eval_overall_accuracy": 0.9585749792874897,
+      "eval_overall_f1": 0.7896592244418331,
+      "eval_overall_precision": 0.7533632286995515,
+      "eval_overall_recall": 0.8296296296296296,
+      "eval_runtime": 0.2742,
+      "eval_samples_per_second": 620.066,
+      "eval_steps_per_second": 10.942,
       "step": 672
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.9401953816413879,
+      "grad_norm": 0.5443127751350403,
       "learning_rate": 4.600000000000001e-05,
-      "loss": 0.1659,
+      "loss": 0.1785,
       "step": 768
     },
     {
       "epoch": 8.0,
-      "eval_LOCATION_f1": 0.7980295566502462,
+      "eval_LOCATION_f1": 0.797979797979798,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.81,
-      "eval_LOCATION_recall": 0.7864077669902912,
-      "eval_ORGANIZATION_f1": 0.7899159663865545,
+      "eval_LOCATION_precision": 0.8315789473684211,
+      "eval_LOCATION_recall": 0.7669902912621359,
+      "eval_ORGANIZATION_f1": 0.7912087912087912,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.7580645161290323,
-      "eval_ORGANIZATION_recall": 0.8245614035087719,
-      "eval_PERSON_f1": 0.9328358208955224,
+      "eval_ORGANIZATION_precision": 0.7461139896373057,
+      "eval_ORGANIZATION_recall": 0.8421052631578947,
+      "eval_PERSON_f1": 0.929889298892989,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9124087591240876,
-      "eval_PERSON_recall": 0.9541984732824428,
-      "eval_loss": 0.12766310572624207,
+      "eval_PERSON_precision": 0.9,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.1342812180519104,
       "eval_overall_accuracy": 0.9652029826014913,
-      "eval_overall_f1": 0.8381642512077295,
-      "eval_overall_precision": 0.8203309692671394,
-      "eval_overall_recall": 0.8567901234567902,
-      "eval_runtime": 0.5896,
-      "eval_samples_per_second": 288.354,
-      "eval_steps_per_second": 5.089,
+      "eval_overall_f1": 0.8379351740696278,
+      "eval_overall_precision": 0.8154205607476636,
+      "eval_overall_recall": 0.8617283950617284,
+      "eval_runtime": 0.2747,
+      "eval_samples_per_second": 618.771,
+      "eval_steps_per_second": 10.919,
       "step": 768
     },
     {
       "epoch": 9.0,
-      "grad_norm": 1.0895642042160034,
+      "grad_norm": 0.889248788356781,
       "learning_rate": 4.55e-05,
-      "loss": 0.1495,
+      "loss": 0.1541,
       "step": 864
     },
     {
       "epoch": 9.0,
-      "eval_LOCATION_f1": 0.835820895522388,
+      "eval_LOCATION_f1": 0.8217821782178217,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8571428571428571,
-      "eval_LOCATION_recall": 0.8155339805825242,
-      "eval_ORGANIZATION_f1": 0.8099173553719009,
+      "eval_LOCATION_precision": 0.8383838383838383,
+      "eval_LOCATION_recall": 0.8058252427184466,
+      "eval_ORGANIZATION_f1": 0.814404432132964,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.765625,
+      "eval_ORGANIZATION_precision": 0.7736842105263158,
       "eval_ORGANIZATION_recall": 0.8596491228070176,
-      "eval_PERSON_f1": 0.9338235294117646,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.900709219858156,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.11188717931509018,
-      "eval_overall_accuracy": 0.9690693178679922,
-      "eval_overall_f1": 0.8564593301435406,
-      "eval_overall_precision": 0.8306264501160093,
-      "eval_overall_recall": 0.8839506172839506,
-      "eval_runtime": 0.6012,
-      "eval_samples_per_second": 282.766,
-      "eval_steps_per_second": 4.99,
+      "eval_PERSON_f1": 0.9264705882352942,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.8936170212765957,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.11753819137811661,
+      "eval_overall_accuracy": 0.968240817453742,
+      "eval_overall_f1": 0.852694610778443,
+      "eval_overall_precision": 0.827906976744186,
+      "eval_overall_recall": 0.8790123456790123,
+      "eval_runtime": 0.2808,
+      "eval_samples_per_second": 605.329,
+      "eval_steps_per_second": 10.682,
       "step": 864
     },
     {
       "epoch": 10.0,
-      "grad_norm": 1.217275857925415,
+      "grad_norm": 0.8680849075317383,
       "learning_rate": 4.5e-05,
-      "loss": 0.1342,
+      "loss": 0.1387,
       "step": 960
     },
     {
       "epoch": 10.0,
-      "eval_LOCATION_f1": 0.8374384236453202,
+      "eval_LOCATION_f1": 0.8195121951219512,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.85,
-      "eval_LOCATION_recall": 0.8252427184466019,
-      "eval_ORGANIZATION_f1": 0.8277777777777778,
+      "eval_LOCATION_precision": 0.8235294117647058,
+      "eval_LOCATION_recall": 0.8155339805825242,
+      "eval_ORGANIZATION_f1": 0.8287292817679557,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.7883597883597884,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
-      "eval_PERSON_f1": 0.929889298892989,
-      "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9,
-      "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.10086647421121597,
-      "eval_overall_accuracy": 0.9707263186964927,
-      "eval_overall_f1": 0.8633093525179856,
-      "eval_overall_precision": 0.8391608391608392,
-      "eval_overall_recall": 0.8888888888888888,
-      "eval_runtime": 0.6119,
-      "eval_samples_per_second": 277.843,
-      "eval_steps_per_second": 4.903,
+      "eval_ORGANIZATION_precision": 0.7853403141361257,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9304029304029305,
+      "eval_PERSON_number": 131,
+      "eval_PERSON_precision": 0.8943661971830986,
+      "eval_PERSON_recall": 0.9694656488549618,
+      "eval_loss": 0.10947263240814209,
+      "eval_overall_accuracy": 0.9696216514774924,
+      "eval_overall_f1": 0.8595238095238096,
+      "eval_overall_precision": 0.8298850574712644,
+      "eval_overall_recall": 0.891358024691358,
+      "eval_runtime": 0.2731,
+      "eval_samples_per_second": 622.547,
+      "eval_steps_per_second": 10.986,
       "step": 960
     },
     {
       "epoch": 11.0,
-      "grad_norm": 2.0349202156066895,
+      "grad_norm": 0.9291552901268005,
       "learning_rate": 4.4500000000000004e-05,
-      "loss": 0.1241,
+      "loss": 0.1275,
       "step": 1056
     },
     {
       "epoch": 11.0,
-      "eval_LOCATION_f1": 0.8365384615384616,
+      "eval_LOCATION_f1": 0.8374384236453202,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8285714285714286,
-      "eval_LOCATION_recall": 0.8446601941747572,
-      "eval_ORGANIZATION_f1": 0.8342857142857143,
+      "eval_LOCATION_precision": 0.85,
+      "eval_LOCATION_recall": 0.8252427184466019,
+      "eval_ORGANIZATION_f1": 0.8333333333333334,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8156424581005587,
-      "eval_ORGANIZATION_recall": 0.8538011695906432,
-      "eval_PERSON_f1": 0.9438202247191011,
+      "eval_ORGANIZATION_precision": 0.7936507936507936,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.929889298892989,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9264705882352942,
+      "eval_PERSON_precision": 0.9,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.09149681031703949,
-      "eval_overall_accuracy": 0.9726594863297432,
-      "eval_overall_f1": 0.8703030303030304,
-      "eval_overall_precision": 0.8547619047619047,
-      "eval_overall_recall": 0.8864197530864197,
-      "eval_runtime": 0.5981,
-      "eval_samples_per_second": 284.231,
-      "eval_steps_per_second": 5.016,
+      "eval_loss": 0.09946061670780182,
+      "eval_overall_accuracy": 0.9710024855012428,
+      "eval_overall_f1": 0.8657074340527577,
+      "eval_overall_precision": 0.8414918414918415,
+      "eval_overall_recall": 0.891358024691358,
+      "eval_runtime": 0.277,
+      "eval_samples_per_second": 613.762,
+      "eval_steps_per_second": 10.831,
       "step": 1056
     },
     {
       "epoch": 12.0,
-      "grad_norm": 3.014051914215088,
+      "grad_norm": 0.6819791793823242,
       "learning_rate": 4.4000000000000006e-05,
-      "loss": 0.1188,
+      "loss": 0.1212,
       "step": 1152
     },
     {
       "epoch": 12.0,
-      "eval_LOCATION_f1": 0.8309178743961353,
+      "eval_LOCATION_f1": 0.8640776699029126,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8269230769230769,
-      "eval_LOCATION_recall": 0.8349514563106796,
-      "eval_ORGANIZATION_f1": 0.8579545454545455,
+      "eval_LOCATION_precision": 0.8640776699029126,
+      "eval_LOCATION_recall": 0.8640776699029126,
+      "eval_ORGANIZATION_f1": 0.837465564738292,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8342541436464088,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
-      "eval_PERSON_f1": 0.9402985074626866,
+      "eval_ORGANIZATION_precision": 0.7916666666666666,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
+      "eval_PERSON_f1": 0.929889298892989,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9197080291970803,
+      "eval_PERSON_precision": 0.9,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.08539355546236038,
-      "eval_overall_accuracy": 0.9745926539629937,
-      "eval_overall_f1": 0.8778718258766627,
-      "eval_overall_precision": 0.8601895734597157,
-      "eval_overall_recall": 0.8962962962962963,
-      "eval_runtime": 0.5994,
-      "eval_samples_per_second": 283.636,
-      "eval_steps_per_second": 5.005,
+      "eval_loss": 0.09351213276386261,
+      "eval_overall_accuracy": 0.9723833195249931,
+      "eval_overall_f1": 0.8738095238095238,
+      "eval_overall_precision": 0.8436781609195402,
+      "eval_overall_recall": 0.9061728395061729,
+      "eval_runtime": 0.2767,
+      "eval_samples_per_second": 614.42,
+      "eval_steps_per_second": 10.843,
       "step": 1152
     },
     {
       "epoch": 13.0,
-      "grad_norm": 2.8539180755615234,
+      "grad_norm": 1.4507771730422974,
       "learning_rate": 4.35e-05,
-      "loss": 0.1102,
+      "loss": 0.1164,
       "step": 1248
     },
     {
       "epoch": 13.0,
-      "eval_LOCATION_f1": 0.8695652173913043,
+      "eval_LOCATION_f1": 0.8585365853658536,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8653846153846154,
-      "eval_LOCATION_recall": 0.8737864077669902,
-      "eval_ORGANIZATION_f1": 0.866096866096866,
+      "eval_LOCATION_precision": 0.8627450980392157,
+      "eval_LOCATION_recall": 0.8543689320388349,
+      "eval_ORGANIZATION_f1": 0.845303867403315,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8444444444444444,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9473684210526315,
+      "eval_ORGANIZATION_precision": 0.8010471204188482,
+      "eval_ORGANIZATION_recall": 0.8947368421052632,
+      "eval_PERSON_f1": 0.929889298892989,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9333333333333333,
+      "eval_PERSON_precision": 0.9,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.07981602847576141,
-      "eval_overall_accuracy": 0.9762496547914941,
-      "eval_overall_f1": 0.8932038834951457,
-      "eval_overall_precision": 0.8782816229116945,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.598,
-      "eval_samples_per_second": 284.265,
-      "eval_steps_per_second": 5.016,
+      "eval_loss": 0.08751274645328522,
+      "eval_overall_accuracy": 0.9723833195249931,
+      "eval_overall_f1": 0.8758949880668259,
+      "eval_overall_precision": 0.8475750577367206,
+      "eval_overall_recall": 0.9061728395061729,
+      "eval_runtime": 0.2748,
+      "eval_samples_per_second": 618.687,
+      "eval_steps_per_second": 10.918,
       "step": 1248
     },
     {
       "epoch": 14.0,
-      "grad_norm": 0.624473512172699,
+      "grad_norm": 0.5297168493270874,
       "learning_rate": 4.3e-05,
-      "loss": 0.1044,
+      "loss": 0.1105,
       "step": 1344
     },
     {
       "epoch": 14.0,
-      "eval_LOCATION_f1": 0.8768472906403941,
+      "eval_LOCATION_f1": 0.8878048780487806,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.89,
-      "eval_LOCATION_recall": 0.8640776699029126,
-      "eval_ORGANIZATION_f1": 0.8830409356725146,
+      "eval_LOCATION_precision": 0.8921568627450981,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8587896253602305,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8830409356725146,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
-      "eval_PERSON_f1": 0.9509433962264152,
+      "eval_ORGANIZATION_precision": 0.8465909090909091,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_PERSON_f1": 0.9438202247191011,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_precision": 0.9264705882352942,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.07802519202232361,
-      "eval_overall_accuracy": 0.9781828224247445,
-      "eval_overall_f1": 0.9037037037037037,
-      "eval_overall_precision": 0.9037037037037037,
+      "eval_loss": 0.08204468339681625,
+      "eval_overall_accuracy": 0.9768019884009942,
+      "eval_overall_f1": 0.8937728937728937,
+      "eval_overall_precision": 0.8840579710144928,
       "eval_overall_recall": 0.9037037037037037,
-      "eval_runtime": 0.5951,
-      "eval_samples_per_second": 285.656,
-      "eval_steps_per_second": 5.041,
+      "eval_runtime": 0.2712,
+      "eval_samples_per_second": 626.816,
+      "eval_steps_per_second": 11.061,
       "step": 1344
     },
     {
       "epoch": 15.0,
-      "grad_norm": 0.996662437915802,
+      "grad_norm": 0.8479740023612976,
       "learning_rate": 4.25e-05,
-      "loss": 0.1009,
+      "loss": 0.1063,
       "step": 1440
     },
     {
       "epoch": 15.0,
-      "eval_LOCATION_f1": 0.891089108910891,
+      "eval_LOCATION_f1": 0.89,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9090909090909091,
-      "eval_LOCATION_recall": 0.8737864077669902,
-      "eval_ORGANIZATION_f1": 0.8788732394366197,
+      "eval_LOCATION_precision": 0.9175257731958762,
+      "eval_LOCATION_recall": 0.8640776699029126,
+      "eval_ORGANIZATION_f1": 0.8446866485013624,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8478260869565217,
-      "eval_ORGANIZATION_recall": 0.9122807017543859,
-      "eval_PERSON_f1": 0.9584905660377357,
+      "eval_ORGANIZATION_precision": 0.7908163265306123,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
+      "eval_PERSON_f1": 0.9333333333333332,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9477611940298507,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.07211567461490631,
-      "eval_overall_accuracy": 0.9781828224247445,
-      "eval_overall_f1": 0.9075425790754258,
-      "eval_overall_precision": 0.894484412470024,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.5925,
-      "eval_samples_per_second": 286.903,
-      "eval_steps_per_second": 5.063,
+      "eval_PERSON_precision": 0.9064748201438849,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.07925137132406235,
+      "eval_overall_accuracy": 0.9751449875724938,
+      "eval_overall_f1": 0.8841099163679809,
+      "eval_overall_precision": 0.8564814814814815,
+      "eval_overall_recall": 0.9135802469135802,
+      "eval_runtime": 0.2741,
+      "eval_samples_per_second": 620.179,
+      "eval_steps_per_second": 10.944,
       "step": 1440
     },
     {
       "epoch": 16.0,
-      "grad_norm": 1.0493874549865723,
+      "grad_norm": 1.6623704433441162,
       "learning_rate": 4.2e-05,
-      "loss": 0.0978,
+      "loss": 0.1018,
       "step": 1536
     },
     {
       "epoch": 16.0,
-      "eval_LOCATION_f1": 0.8932038834951457,
+      "eval_LOCATION_f1": 0.8846153846153847,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8932038834951457,
+      "eval_LOCATION_precision": 0.8761904761904762,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8725212464589235,
+      "eval_ORGANIZATION_f1": 0.8555858310626703,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8461538461538461,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
-      "eval_PERSON_f1": 0.9509433962264152,
+      "eval_ORGANIZATION_precision": 0.8010204081632653,
+      "eval_ORGANIZATION_recall": 0.9181286549707602,
+      "eval_PERSON_f1": 0.9333333333333332,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_precision": 0.9064748201438849,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.06955698877573013,
-      "eval_overall_accuracy": 0.9779066556199945,
-      "eval_overall_f1": 0.9029126213592233,
-      "eval_overall_precision": 0.8878281622911695,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.6569,
-      "eval_samples_per_second": 258.806,
-      "eval_steps_per_second": 4.567,
+      "eval_loss": 0.07826149463653564,
+      "eval_overall_accuracy": 0.9748688207677437,
+      "eval_overall_f1": 0.8875739644970415,
+      "eval_overall_precision": 0.8522727272727273,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2734,
+      "eval_samples_per_second": 621.692,
+      "eval_steps_per_second": 10.971,
       "step": 1536
     },
     {
       "epoch": 17.0,
-      "grad_norm": 0.8617897629737854,
+      "grad_norm": 0.8166274428367615,
       "learning_rate": 4.15e-05,
-      "loss": 0.0962,
+      "loss": 0.0986,
       "step": 1632
     },
     {
       "epoch": 17.0,
-      "eval_LOCATION_f1": 0.9064039408866995,
+      "eval_LOCATION_f1": 0.9019607843137256,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.92,
+      "eval_LOCATION_precision": 0.9108910891089109,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8869565217391304,
+      "eval_ORGANIZATION_f1": 0.8627450980392157,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8793103448275862,
-      "eval_ORGANIZATION_recall": 0.8947368421052632,
-      "eval_PERSON_f1": 0.962121212121212,
+      "eval_ORGANIZATION_precision": 0.8279569892473119,
+      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_PERSON_f1": 0.9548872180451129,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9548872180451128,
+      "eval_PERSON_precision": 0.9407407407407408,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.06796535849571228,
-      "eval_overall_accuracy": 0.980115990057995,
-      "eval_overall_f1": 0.916256157635468,
-      "eval_overall_precision": 0.914004914004914,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.6013,
-      "eval_samples_per_second": 282.726,
-      "eval_steps_per_second": 4.989,
+      "eval_loss": 0.07254834473133087,
+      "eval_overall_accuracy": 0.9779066556199945,
+      "eval_overall_f1": 0.9020556227327691,
+      "eval_overall_precision": 0.8838862559241706,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.274,
+      "eval_samples_per_second": 620.343,
+      "eval_steps_per_second": 10.947,
       "step": 1632
     },
     {
       "epoch": 18.0,
-      "grad_norm": 0.8590739965438843,
+      "grad_norm": 0.7182817459106445,
       "learning_rate": 4.1e-05,
-      "loss": 0.0909,
+      "loss": 0.093,
       "step": 1728
     },
     {
       "epoch": 18.0,
-      "eval_LOCATION_f1": 0.896551724137931,
+      "eval_LOCATION_f1": 0.8921568627450982,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.91,
+      "eval_LOCATION_precision": 0.900990099009901,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8831908831908832,
+      "eval_ORGANIZATION_f1": 0.8764044943820224,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8611111111111112,
-      "eval_ORGANIZATION_recall": 0.9064327485380117,
-      "eval_PERSON_f1": 0.9509433962264152,
+      "eval_ORGANIZATION_precision": 0.8432432432432433,
+      "eval_ORGANIZATION_recall": 0.9122807017543859,
+      "eval_PERSON_f1": 0.9473684210526315,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_precision": 0.9333333333333333,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.0644010454416275,
-      "eval_overall_accuracy": 0.9798398232532449,
-      "eval_overall_f1": 0.9084249084249084,
-      "eval_overall_precision": 0.8985507246376812,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.5942,
-      "eval_samples_per_second": 286.117,
-      "eval_steps_per_second": 5.049,
+      "eval_loss": 0.06929948180913925,
+      "eval_overall_accuracy": 0.9779066556199945,
+      "eval_overall_f1": 0.9031476997578692,
+      "eval_overall_precision": 0.8859857482185273,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 623.704,
+      "eval_steps_per_second": 11.007,
       "step": 1728
     },
     {
       "epoch": 19.0,
-      "grad_norm": 2.493396520614624,
+      "grad_norm": 1.4325016736984253,
       "learning_rate": 4.05e-05,
-      "loss": 0.088,
+      "loss": 0.0897,
       "step": 1824
     },
     {
       "epoch": 19.0,
-      "eval_LOCATION_f1": 0.9009900990099011,
+      "eval_LOCATION_f1": 0.8846153846153847,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9191919191919192,
-      "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8908045977011494,
+      "eval_LOCATION_precision": 0.8761904761904762,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.8757062146892655,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8757062146892656,
+      "eval_ORGANIZATION_precision": 0.8469945355191257,
       "eval_ORGANIZATION_recall": 0.9064327485380117,
-      "eval_PERSON_f1": 0.9545454545454546,
+      "eval_PERSON_f1": 0.9477611940298507,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9473684210526315,
-      "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.0634172335267067,
-      "eval_overall_accuracy": 0.9806683236674951,
-      "eval_overall_f1": 0.914004914004914,
-      "eval_overall_precision": 0.9095354523227384,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.5943,
-      "eval_samples_per_second": 286.044,
-      "eval_steps_per_second": 5.048,
+      "eval_PERSON_precision": 0.927007299270073,
+      "eval_PERSON_recall": 0.9694656488549618,
+      "eval_loss": 0.06990908831357956,
+      "eval_overall_accuracy": 0.9781828224247445,
+      "eval_overall_f1": 0.9012048192771085,
+      "eval_overall_precision": 0.88,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2715,
+      "eval_samples_per_second": 626.106,
+      "eval_steps_per_second": 11.049,
       "step": 1824
     },
     {
       "epoch": 20.0,
-      "grad_norm": 0.5532075762748718,
+      "grad_norm": 0.7544979453086853,
       "learning_rate": 4e-05,
-      "loss": 0.0836,
+      "loss": 0.0876,
       "step": 1920
     },
     {
@@ -610,153 +610,153 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.8846153846153846,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8863636363636364,
+      "eval_ORGANIZATION_f1": 0.8611111111111113,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.861878453038674,
-      "eval_ORGANIZATION_recall": 0.9122807017543859,
-      "eval_PERSON_f1": 0.9509433962264152,
+      "eval_ORGANIZATION_precision": 0.8201058201058201,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
+      "eval_PERSON_f1": 0.9333333333333332,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_precision": 0.9064748201438849,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.06357116252183914,
-      "eval_overall_accuracy": 0.9798398232532449,
-      "eval_overall_f1": 0.9077669902912622,
-      "eval_overall_precision": 0.8926014319809069,
-      "eval_overall_recall": 0.9234567901234568,
-      "eval_runtime": 0.5958,
-      "eval_samples_per_second": 285.346,
-      "eval_steps_per_second": 5.036,
+      "eval_loss": 0.06785619258880615,
+      "eval_overall_accuracy": 0.9765258215962441,
+      "eval_overall_f1": 0.8912783751493429,
+      "eval_overall_precision": 0.8634259259259259,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2728,
+      "eval_samples_per_second": 623.147,
+      "eval_steps_per_second": 10.997,
       "step": 1920
     },
     {
       "epoch": 21.0,
-      "grad_norm": 1.6820136308670044,
+      "grad_norm": 0.9009093046188354,
       "learning_rate": 3.9500000000000005e-05,
-      "loss": 0.0836,
+      "loss": 0.0846,
       "step": 2016
     },
     {
       "epoch": 21.0,
-      "eval_LOCATION_f1": 0.8888888888888888,
+      "eval_LOCATION_f1": 0.8803827751196173,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8846153846153846,
+      "eval_LOCATION_precision": 0.8679245283018868,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8760806916426512,
+      "eval_ORGANIZATION_f1": 0.8707865168539327,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8636363636363636,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9545454545454546,
+      "eval_ORGANIZATION_precision": 0.8378378378378378,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
+      "eval_PERSON_f1": 0.9402985074626866,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_precision": 0.9197080291970803,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05993271991610527,
-      "eval_overall_accuracy": 0.9806683236674951,
-      "eval_overall_f1": 0.9046454767726161,
-      "eval_overall_precision": 0.8958837772397095,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.5978,
-      "eval_samples_per_second": 284.362,
-      "eval_steps_per_second": 5.018,
+      "eval_loss": 0.06543691456317902,
+      "eval_overall_accuracy": 0.9784589892294946,
+      "eval_overall_f1": 0.8955582232893157,
+      "eval_overall_precision": 0.8714953271028038,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2743,
+      "eval_samples_per_second": 619.702,
+      "eval_steps_per_second": 10.936,
       "step": 2016
     },
     {
       "epoch": 22.0,
-      "grad_norm": 0.9439449906349182,
+      "grad_norm": 1.1362422704696655,
       "learning_rate": 3.9000000000000006e-05,
-      "loss": 0.0827,
+      "loss": 0.0843,
       "step": 2112
     },
     {
       "epoch": 22.0,
-      "eval_LOCATION_f1": 0.8761904761904761,
+      "eval_LOCATION_f1": 0.8932038834951457,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8598130841121495,
+      "eval_LOCATION_precision": 0.8932038834951457,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8757062146892655,
+      "eval_ORGANIZATION_f1": 0.8784530386740331,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8469945355191257,
-      "eval_ORGANIZATION_recall": 0.9064327485380117,
-      "eval_PERSON_f1": 0.9509433962264152,
+      "eval_ORGANIZATION_precision": 0.8324607329842932,
+      "eval_ORGANIZATION_recall": 0.9298245614035088,
+      "eval_PERSON_f1": 0.9402985074626866,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_precision": 0.9197080291970803,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05957403779029846,
-      "eval_overall_accuracy": 0.9798398232532449,
-      "eval_overall_f1": 0.8998793727382389,
-      "eval_overall_precision": 0.8797169811320755,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.5967,
-      "eval_samples_per_second": 284.901,
-      "eval_steps_per_second": 5.028,
+      "eval_loss": 0.06642135977745056,
+      "eval_overall_accuracy": 0.9787351560342447,
+      "eval_overall_f1": 0.9019138755980862,
+      "eval_overall_precision": 0.8747099767981439,
+      "eval_overall_recall": 0.9308641975308642,
+      "eval_runtime": 0.2703,
+      "eval_samples_per_second": 628.899,
+      "eval_steps_per_second": 11.098,
       "step": 2112
     },
     {
       "epoch": 23.0,
-      "grad_norm": 1.140651822090149,
+      "grad_norm": 1.139490008354187,
       "learning_rate": 3.85e-05,
-      "loss": 0.077,
+      "loss": 0.0823,
       "step": 2208
     },
     {
       "epoch": 23.0,
-      "eval_LOCATION_f1": 0.8750000000000001,
+      "eval_LOCATION_f1": 0.8803827751196173,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8666666666666667,
-      "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8818443804034583,
+      "eval_LOCATION_precision": 0.8679245283018868,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.8685714285714285,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8693181818181818,
-      "eval_ORGANIZATION_recall": 0.8947368421052632,
+      "eval_ORGANIZATION_precision": 0.8491620111731844,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.058286890387535095,
-      "eval_overall_accuracy": 0.9806683236674951,
-      "eval_overall_f1": 0.9070904645476773,
-      "eval_overall_precision": 0.8983050847457628,
+      "eval_loss": 0.06113022193312645,
+      "eval_overall_accuracy": 0.980115990057995,
+      "eval_overall_f1": 0.902676399026764,
+      "eval_overall_precision": 0.8896882494004796,
       "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.6001,
-      "eval_samples_per_second": 283.301,
-      "eval_steps_per_second": 4.999,
+      "eval_runtime": 0.2712,
+      "eval_samples_per_second": 626.885,
+      "eval_steps_per_second": 11.063,
       "step": 2208
     },
     {
       "epoch": 24.0,
-      "grad_norm": 0.9434022307395935,
+      "grad_norm": 0.9067116975784302,
       "learning_rate": 3.8e-05,
-      "loss": 0.0774,
+      "loss": 0.0808,
       "step": 2304
     },
     {
       "epoch": 24.0,
-      "eval_LOCATION_f1": 0.861111111111111,
+      "eval_LOCATION_f1": 0.8666666666666667,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8230088495575221,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8529411764705882,
+      "eval_LOCATION_precision": 0.8504672897196262,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8700564971751413,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8579881656804734,
-      "eval_ORGANIZATION_recall": 0.847953216374269,
-      "eval_PERSON_f1": 0.9545454545454546,
+      "eval_ORGANIZATION_precision": 0.8415300546448088,
+      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_PERSON_f1": 0.962121212121212,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9473684210526315,
-      "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05800532177090645,
-      "eval_overall_accuracy": 0.9787351560342447,
-      "eval_overall_f1": 0.8878048780487805,
-      "eval_overall_precision": 0.8771084337349397,
-      "eval_overall_recall": 0.8987654320987655,
-      "eval_runtime": 0.6032,
-      "eval_samples_per_second": 281.813,
-      "eval_steps_per_second": 4.973,
+      "eval_PERSON_precision": 0.9548872180451128,
+      "eval_PERSON_recall": 0.9694656488549618,
+      "eval_loss": 0.0627431645989418,
+      "eval_overall_accuracy": 0.9798398232532449,
+      "eval_overall_f1": 0.8985507246376813,
+      "eval_overall_precision": 0.8794326241134752,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2724,
+      "eval_samples_per_second": 624.125,
+      "eval_steps_per_second": 11.014,
       "step": 2304
     },
     {
       "epoch": 25.0,
-      "grad_norm": 1.0291742086410522,
+      "grad_norm": 1.520267128944397,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0782,
+      "loss": 0.0809,
       "step": 2400
     },
     {
@@ -765,60 +765,60 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.875,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8816568047337278,
+      "eval_ORGANIZATION_f1": 0.8732394366197183,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8922155688622755,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.842391304347826,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
       "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9473684210526315,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05705862119793892,
-      "eval_overall_accuracy": 0.9798398232532449,
-      "eval_overall_f1": 0.9048207663782447,
-      "eval_overall_precision": 0.905940594059406,
-      "eval_overall_recall": 0.9037037037037037,
-      "eval_runtime": 0.6018,
-      "eval_samples_per_second": 282.474,
-      "eval_steps_per_second": 4.985,
+      "eval_loss": 0.05983828753232956,
+      "eval_overall_accuracy": 0.9806683236674951,
+      "eval_overall_f1": 0.9007263922518159,
+      "eval_overall_precision": 0.8836104513064132,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.272,
+      "eval_samples_per_second": 624.952,
+      "eval_steps_per_second": 11.029,
       "step": 2400
     },
     {
       "epoch": 26.0,
-      "grad_norm": 0.6928281784057617,
+      "grad_norm": 0.6365208029747009,
       "learning_rate": 3.7e-05,
-      "loss": 0.0771,
+      "loss": 0.078,
       "step": 2496
     },
     {
       "epoch": 26.0,
-      "eval_LOCATION_f1": 0.8975609756097561,
+      "eval_LOCATION_f1": 0.8921568627450982,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9019607843137255,
-      "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8977272727272728,
+      "eval_LOCATION_precision": 0.900990099009901,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8795518207282913,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8729281767955801,
-      "eval_ORGANIZATION_recall": 0.9239766081871345,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8440860215053764,
+      "eval_ORGANIZATION_recall": 0.9181286549707602,
+      "eval_PERSON_f1": 0.9509433962264152,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.055970560759305954,
-      "eval_overall_accuracy": 0.9820491576912456,
-      "eval_overall_f1": 0.9195121951219513,
-      "eval_overall_precision": 0.908433734939759,
-      "eval_overall_recall": 0.9308641975308642,
-      "eval_runtime": 0.6049,
-      "eval_samples_per_second": 281.02,
-      "eval_steps_per_second": 4.959,
+      "eval_PERSON_precision": 0.9402985074626866,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.058052726089954376,
+      "eval_overall_accuracy": 0.9817729908864954,
+      "eval_overall_f1": 0.9055690072639224,
+      "eval_overall_precision": 0.8883610451306413,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.273,
+      "eval_samples_per_second": 622.679,
+      "eval_steps_per_second": 10.988,
       "step": 2496
     },
     {
       "epoch": 27.0,
-      "grad_norm": 1.4192240238189697,
+      "grad_norm": 1.504185676574707,
       "learning_rate": 3.65e-05,
-      "loss": 0.0726,
+      "loss": 0.0774,
       "step": 2592
     },
     {
@@ -827,215 +827,215 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.92,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.9008498583569404,
+      "eval_ORGANIZATION_f1": 0.888268156424581,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8736263736263736,
+      "eval_ORGANIZATION_precision": 0.8502673796791443,
       "eval_ORGANIZATION_recall": 0.9298245614035088,
-      "eval_PERSON_f1": 0.9545454545454546,
+      "eval_PERSON_f1": 0.9509433962264152,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_precision": 0.9402985074626866,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05344228819012642,
-      "eval_overall_accuracy": 0.9826014913007457,
-      "eval_overall_f1": 0.9195121951219513,
-      "eval_overall_precision": 0.908433734939759,
+      "eval_loss": 0.058188218623399734,
+      "eval_overall_accuracy": 0.9812206572769953,
+      "eval_overall_f1": 0.9128329297820823,
+      "eval_overall_precision": 0.8954869358669834,
       "eval_overall_recall": 0.9308641975308642,
-      "eval_runtime": 0.605,
-      "eval_samples_per_second": 281.0,
-      "eval_steps_per_second": 4.959,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 623.662,
+      "eval_steps_per_second": 11.006,
       "step": 2592
     },
     {
       "epoch": 28.0,
-      "grad_norm": 1.0541571378707886,
+      "grad_norm": 0.8313342928886414,
       "learning_rate": 3.6e-05,
-      "loss": 0.0712,
+      "loss": 0.0732,
       "step": 2688
     },
     {
       "epoch": 28.0,
-      "eval_LOCATION_f1": 0.8932038834951457,
+      "eval_LOCATION_f1": 0.8975609756097561,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8932038834951457,
+      "eval_LOCATION_precision": 0.9019607843137255,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8939828080229225,
+      "eval_ORGANIZATION_f1": 0.8857142857142858,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764044943820225,
-      "eval_ORGANIZATION_recall": 0.9122807017543859,
+      "eval_ORGANIZATION_precision": 0.8659217877094972,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
       "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9473684210526315,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05282684788107872,
-      "eval_overall_accuracy": 0.9826014913007457,
-      "eval_overall_f1": 0.9133089133089134,
-      "eval_overall_precision": 0.9033816425120773,
-      "eval_overall_recall": 0.9234567901234568,
-      "eval_runtime": 0.6016,
-      "eval_samples_per_second": 282.587,
-      "eval_steps_per_second": 4.987,
+      "eval_loss": 0.062320366501808167,
+      "eval_overall_accuracy": 0.9814968240817453,
+      "eval_overall_f1": 0.9108669108669109,
+      "eval_overall_precision": 0.9009661835748792,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.273,
+      "eval_samples_per_second": 622.639,
+      "eval_steps_per_second": 10.988,
       "step": 2688
     },
     {
       "epoch": 29.0,
-      "grad_norm": 0.6951781511306763,
+      "grad_norm": 0.6814190149307251,
       "learning_rate": 3.55e-05,
-      "loss": 0.0703,
+      "loss": 0.0746,
       "step": 2784
     },
     {
       "epoch": 29.0,
-      "eval_LOCATION_f1": 0.9009900990099011,
+      "eval_LOCATION_f1": 0.9019607843137256,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9191919191919192,
-      "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.9008498583569404,
+      "eval_LOCATION_precision": 0.9108910891089109,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.9028571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8736263736263736,
-      "eval_ORGANIZATION_recall": 0.9298245614035088,
+      "eval_ORGANIZATION_precision": 0.88268156424581,
+      "eval_ORGANIZATION_recall": 0.9239766081871345,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.05054891109466553,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9217603911980441,
-      "eval_overall_precision": 0.9128329297820823,
+      "eval_loss": 0.05526537075638771,
+      "eval_overall_accuracy": 0.9828776581054958,
+      "eval_overall_f1": 0.922888616891065,
+      "eval_overall_precision": 0.9150485436893204,
       "eval_overall_recall": 0.9308641975308642,
-      "eval_runtime": 0.5939,
-      "eval_samples_per_second": 286.258,
-      "eval_steps_per_second": 5.052,
+      "eval_runtime": 0.2751,
+      "eval_samples_per_second": 617.869,
+      "eval_steps_per_second": 10.904,
       "step": 2784
     },
     {
       "epoch": 30.0,
-      "grad_norm": 0.4239547550678253,
+      "grad_norm": 1.8307719230651855,
       "learning_rate": 3.5e-05,
-      "loss": 0.0649,
+      "loss": 0.0695,
       "step": 2880
     },
     {
       "epoch": 30.0,
-      "eval_LOCATION_f1": 0.9054726368159205,
+      "eval_LOCATION_f1": 0.9019607843137256,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9285714285714286,
-      "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.9065155807365438,
+      "eval_LOCATION_precision": 0.9108910891089109,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.9028571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8791208791208791,
-      "eval_ORGANIZATION_recall": 0.935672514619883,
+      "eval_ORGANIZATION_precision": 0.88268156424581,
+      "eval_ORGANIZATION_recall": 0.9239766081871345,
       "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9473684210526315,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05008954927325249,
-      "eval_overall_accuracy": 0.983706158519746,
-      "eval_overall_f1": 0.9217603911980441,
-      "eval_overall_precision": 0.9128329297820823,
-      "eval_overall_recall": 0.9308641975308642,
-      "eval_runtime": 0.606,
-      "eval_samples_per_second": 280.539,
-      "eval_steps_per_second": 4.951,
+      "eval_loss": 0.05361052602529526,
+      "eval_overall_accuracy": 0.9831538249102458,
+      "eval_overall_f1": 0.9193154034229828,
+      "eval_overall_precision": 0.910411622276029,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2712,
+      "eval_samples_per_second": 626.924,
+      "eval_steps_per_second": 11.063,
       "step": 2880
     },
     {
       "epoch": 31.0,
-      "grad_norm": 1.2522234916687012,
+      "grad_norm": 0.5288766026496887,
       "learning_rate": 3.45e-05,
-      "loss": 0.0642,
+      "loss": 0.0691,
       "step": 2976
     },
     {
       "epoch": 31.0,
-      "eval_LOCATION_f1": 0.8720379146919431,
+      "eval_LOCATION_f1": 0.8846153846153847,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8518518518518519,
+      "eval_LOCATION_precision": 0.8761904761904762,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8786127167630058,
+      "eval_ORGANIZATION_f1": 0.893371757925072,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8685714285714285,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8806818181818182,
+      "eval_ORGANIZATION_recall": 0.9064327485380117,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.05007793754339218,
-      "eval_overall_accuracy": 0.9820491576912456,
-      "eval_overall_f1": 0.9048780487804877,
-      "eval_overall_precision": 0.8939759036144578,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.5983,
-      "eval_samples_per_second": 284.148,
-      "eval_steps_per_second": 5.014,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.05325145646929741,
+      "eval_overall_accuracy": 0.9826014913007457,
+      "eval_overall_f1": 0.9108669108669109,
+      "eval_overall_precision": 0.9009661835748792,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2729,
+      "eval_samples_per_second": 622.963,
+      "eval_steps_per_second": 10.993,
       "step": 2976
     },
     {
       "epoch": 32.0,
-      "grad_norm": 0.8345080018043518,
+      "grad_norm": 0.7085039019584656,
       "learning_rate": 3.4000000000000007e-05,
-      "loss": 0.0664,
+      "loss": 0.0665,
       "step": 3072
     },
     {
       "epoch": 32.0,
-      "eval_LOCATION_f1": 0.8625592417061612,
+      "eval_LOCATION_f1": 0.883495145631068,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8425925925925926,
+      "eval_LOCATION_precision": 0.883495145631068,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8835820895522388,
+      "eval_ORGANIZATION_f1": 0.8825214899713466,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9024390243902439,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.8651685393258427,
+      "eval_ORGANIZATION_recall": 0.9005847953216374,
       "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9473684210526315,
       "eval_PERSON_recall": 0.9618320610687023,
-      "eval_loss": 0.05058050900697708,
-      "eval_overall_accuracy": 0.9817729908864954,
-      "eval_overall_f1": 0.9012345679012346,
-      "eval_overall_precision": 0.9012345679012346,
-      "eval_overall_recall": 0.9012345679012346,
-      "eval_runtime": 0.6023,
-      "eval_samples_per_second": 282.266,
-      "eval_steps_per_second": 4.981,
+      "eval_loss": 0.05177022144198418,
+      "eval_overall_accuracy": 0.9823253244959956,
+      "eval_overall_f1": 0.905982905982906,
+      "eval_overall_precision": 0.8961352657004831,
+      "eval_overall_recall": 0.9160493827160494,
+      "eval_runtime": 0.2709,
+      "eval_samples_per_second": 627.553,
+      "eval_steps_per_second": 11.074,
       "step": 3072
     },
     {
       "epoch": 33.0,
-      "grad_norm": 0.7092554569244385,
+      "grad_norm": 0.6961383819580078,
       "learning_rate": 3.35e-05,
-      "loss": 0.0659,
+      "loss": 0.0649,
       "step": 3168
     },
     {
       "epoch": 33.0,
-      "eval_LOCATION_f1": 0.8584905660377358,
+      "eval_LOCATION_f1": 0.8598130841121495,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8348623853211009,
-      "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8842729970326408,
+      "eval_LOCATION_precision": 0.8288288288288288,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.8802395209580838,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8975903614457831,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.901840490797546,
+      "eval_ORGANIZATION_recall": 0.8596491228070176,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.0509335920214653,
-      "eval_overall_accuracy": 0.9820491576912456,
-      "eval_overall_f1": 0.9039408866995073,
-      "eval_overall_precision": 0.9017199017199017,
-      "eval_overall_recall": 0.9061728395061729,
-      "eval_runtime": 0.5971,
-      "eval_samples_per_second": 284.704,
-      "eval_steps_per_second": 5.024,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.052676014602184296,
+      "eval_overall_accuracy": 0.9809444904722452,
+      "eval_overall_f1": 0.8990147783251232,
+      "eval_overall_precision": 0.8968058968058968,
+      "eval_overall_recall": 0.9012345679012346,
+      "eval_runtime": 0.273,
+      "eval_samples_per_second": 622.752,
+      "eval_steps_per_second": 10.99,
       "step": 3168
     },
     {
       "epoch": 34.0,
-      "grad_norm": 0.8459410071372986,
+      "grad_norm": 0.3737340569496155,
       "learning_rate": 3.3e-05,
-      "loss": 0.0667,
+      "loss": 0.0645,
       "step": 3264
     },
     {
@@ -1044,277 +1044,277 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.883495145631068,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.8908045977011494,
+      "eval_ORGANIZATION_f1": 0.9043478260869565,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8757062146892656,
-      "eval_ORGANIZATION_recall": 0.9064327485380117,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.896551724137931,
+      "eval_ORGANIZATION_recall": 0.9122807017543859,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.050714682787656784,
-      "eval_overall_accuracy": 0.9826014913007457,
-      "eval_overall_f1": 0.9130966952264382,
-      "eval_overall_precision": 0.9053398058252428,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.05056975781917572,
+      "eval_overall_accuracy": 0.9842584921292461,
+      "eval_overall_f1": 0.9153374233128834,
+      "eval_overall_precision": 0.9097560975609756,
       "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.5977,
-      "eval_samples_per_second": 284.432,
-      "eval_steps_per_second": 5.019,
+      "eval_runtime": 0.2748,
+      "eval_samples_per_second": 618.671,
+      "eval_steps_per_second": 10.918,
       "step": 3264
     },
     {
       "epoch": 35.0,
-      "grad_norm": 1.9158438444137573,
+      "grad_norm": 1.4867804050445557,
       "learning_rate": 3.2500000000000004e-05,
-      "loss": 0.0639,
+      "loss": 0.063,
       "step": 3360
     },
     {
       "epoch": 35.0,
-      "eval_LOCATION_f1": 0.8761904761904761,
+      "eval_LOCATION_f1": 0.8666666666666667,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8598130841121495,
-      "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8934911242603552,
+      "eval_LOCATION_precision": 0.8504672897196262,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8888888888888888,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9041916167664671,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8888888888888888,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.0509558841586113,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.051477495580911636,
       "eval_overall_accuracy": 0.9826014913007457,
-      "eval_overall_f1": 0.9124537607891491,
-      "eval_overall_precision": 0.9113300492610837,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.6048,
-      "eval_samples_per_second": 281.07,
-      "eval_steps_per_second": 4.96,
+      "eval_overall_f1": 0.9044117647058824,
+      "eval_overall_precision": 0.8978102189781022,
+      "eval_overall_recall": 0.9111111111111111,
+      "eval_runtime": 0.2727,
+      "eval_samples_per_second": 623.41,
+      "eval_steps_per_second": 11.001,
       "step": 3360
     },
     {
       "epoch": 36.0,
-      "grad_norm": 1.0001205205917358,
+      "grad_norm": 0.9459262490272522,
       "learning_rate": 3.2000000000000005e-05,
-      "loss": 0.0632,
+      "loss": 0.0637,
       "step": 3456
     },
     {
       "epoch": 36.0,
-      "eval_LOCATION_f1": 0.8846153846153847,
+      "eval_LOCATION_f1": 0.8666666666666667,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8761904761904762,
-      "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.9032258064516129,
+      "eval_LOCATION_precision": 0.8504672897196262,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8830409356725146,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9058823529411765,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8830409356725146,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.050493232905864716,
-      "eval_overall_accuracy": 0.983706158519746,
-      "eval_overall_f1": 0.91871921182266,
-      "eval_overall_precision": 0.9164619164619164,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.5977,
-      "eval_samples_per_second": 284.414,
-      "eval_steps_per_second": 5.019,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.050841640681028366,
+      "eval_overall_accuracy": 0.9817729908864954,
+      "eval_overall_f1": 0.9019607843137256,
+      "eval_overall_precision": 0.8953771289537713,
+      "eval_overall_recall": 0.908641975308642,
+      "eval_runtime": 0.2727,
+      "eval_samples_per_second": 623.352,
+      "eval_steps_per_second": 11.0,
       "step": 3456
     },
     {
       "epoch": 37.0,
-      "grad_norm": 2.2751259803771973,
+      "grad_norm": 2.0613455772399902,
       "learning_rate": 3.15e-05,
-      "loss": 0.0632,
+      "loss": 0.0614,
       "step": 3552
     },
     {
       "epoch": 37.0,
-      "eval_LOCATION_f1": 0.9054726368159205,
+      "eval_LOCATION_f1": 0.8921568627450982,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9285714285714286,
+      "eval_LOCATION_precision": 0.900990099009901,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.9132947976878613,
+      "eval_ORGANIZATION_f1": 0.8977272727272728,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9028571428571428,
+      "eval_ORGANIZATION_precision": 0.8729281767955801,
       "eval_ORGANIZATION_recall": 0.9239766081871345,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04872192069888115,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.928395061728395,
-      "eval_overall_precision": 0.928395061728395,
-      "eval_overall_recall": 0.928395061728395,
-      "eval_runtime": 0.5916,
-      "eval_samples_per_second": 287.366,
-      "eval_steps_per_second": 5.071,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.049506776034832,
+      "eval_overall_accuracy": 0.9828776581054958,
+      "eval_overall_f1": 0.9146341463414634,
+      "eval_overall_precision": 0.9036144578313253,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2728,
+      "eval_samples_per_second": 623.161,
+      "eval_steps_per_second": 10.997,
       "step": 3552
     },
     {
       "epoch": 38.0,
-      "grad_norm": 0.6949302554130554,
+      "grad_norm": 0.5442711710929871,
       "learning_rate": 3.1e-05,
-      "loss": 0.0605,
+      "loss": 0.0599,
       "step": 3648
     },
     {
       "epoch": 38.0,
-      "eval_LOCATION_f1": 0.8803827751196173,
+      "eval_LOCATION_f1": 0.8708133971291866,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8679245283018868,
-      "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8922155688622756,
+      "eval_LOCATION_precision": 0.8584905660377359,
+      "eval_LOCATION_recall": 0.883495145631068,
+      "eval_ORGANIZATION_f1": 0.8875739644970414,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9141104294478528,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8982035928143712,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9545454545454546,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
-      "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.05016703903675079,
-      "eval_overall_accuracy": 0.9831538249102458,
-      "eval_overall_f1": 0.913151364764268,
-      "eval_overall_precision": 0.9177057356608479,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.5914,
-      "eval_samples_per_second": 287.442,
-      "eval_steps_per_second": 5.073,
+      "eval_PERSON_precision": 0.9473684210526315,
+      "eval_PERSON_recall": 0.9618320610687023,
+      "eval_loss": 0.04945772886276245,
+      "eval_overall_accuracy": 0.9820491576912456,
+      "eval_overall_f1": 0.9050554870530209,
+      "eval_overall_precision": 0.9039408866995073,
+      "eval_overall_recall": 0.9061728395061729,
+      "eval_runtime": 0.2707,
+      "eval_samples_per_second": 628.033,
+      "eval_steps_per_second": 11.083,
       "step": 3648
     },
     {
       "epoch": 39.0,
-      "grad_norm": 0.5452874302864075,
+      "grad_norm": 0.5759150385856628,
       "learning_rate": 3.05e-05,
-      "loss": 0.0609,
+      "loss": 0.06,
       "step": 3744
     },
     {
       "epoch": 39.0,
-      "eval_LOCATION_f1": 0.8803827751196173,
+      "eval_LOCATION_f1": 0.8720379146919431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8679245283018868,
+      "eval_LOCATION_precision": 0.8518518518518519,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8953488372093024,
+      "eval_ORGANIZATION_f1": 0.877906976744186,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8901734104046243,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8728323699421965,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.046954866498708725,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.9142156862745098,
-      "eval_overall_precision": 0.9075425790754258,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.6085,
-      "eval_samples_per_second": 279.378,
-      "eval_steps_per_second": 4.93,
+      "eval_loss": 0.0495310053229332,
+      "eval_overall_accuracy": 0.9820491576912456,
+      "eval_overall_f1": 0.9057527539779682,
+      "eval_overall_precision": 0.8980582524271845,
+      "eval_overall_recall": 0.9135802469135802,
+      "eval_runtime": 0.277,
+      "eval_samples_per_second": 613.735,
+      "eval_steps_per_second": 10.831,
       "step": 3744
     },
     {
       "epoch": 40.0,
-      "grad_norm": 2.737624406814575,
+      "grad_norm": 0.95399409532547,
       "learning_rate": 3e-05,
-      "loss": 0.0601,
+      "loss": 0.0576,
       "step": 3840
     },
     {
       "epoch": 40.0,
-      "eval_LOCATION_f1": 0.8878048780487806,
+      "eval_LOCATION_f1": 0.8750000000000001,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8921568627450981,
+      "eval_LOCATION_precision": 0.8666666666666667,
       "eval_LOCATION_recall": 0.883495145631068,
-      "eval_ORGANIZATION_f1": 0.9032258064516129,
+      "eval_ORGANIZATION_f1": 0.8941176470588235,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9058823529411765,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_ORGANIZATION_precision": 0.8994082840236687,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.046938493847846985,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.919653893695921,
-      "eval_overall_precision": 0.9207920792079208,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.5981,
-      "eval_samples_per_second": 284.253,
-      "eval_steps_per_second": 5.016,
+      "eval_loss": 0.04800355061888695,
+      "eval_overall_accuracy": 0.983706158519746,
+      "eval_overall_f1": 0.9124537607891491,
+      "eval_overall_precision": 0.9113300492610837,
+      "eval_overall_recall": 0.9135802469135802,
+      "eval_runtime": 0.2713,
+      "eval_samples_per_second": 626.595,
+      "eval_steps_per_second": 11.058,
       "step": 3840
     },
     {
       "epoch": 41.0,
-      "grad_norm": 0.9197700619697571,
+      "grad_norm": 0.808504581451416,
       "learning_rate": 2.95e-05,
-      "loss": 0.0588,
+      "loss": 0.0597,
       "step": 3936
     },
     {
       "epoch": 41.0,
-      "eval_LOCATION_f1": 0.9064039408866995,
+      "eval_LOCATION_f1": 0.8803827751196173,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.92,
+      "eval_LOCATION_precision": 0.8679245283018868,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.9137931034482759,
+      "eval_ORGANIZATION_f1": 0.8876080691642652,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8983050847457628,
-      "eval_ORGANIZATION_recall": 0.9298245614035088,
+      "eval_ORGANIZATION_precision": 0.875,
+      "eval_ORGANIZATION_recall": 0.9005847953216374,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04637211561203003,
-      "eval_overall_accuracy": 0.9864678265672466,
-      "eval_overall_f1": 0.9287469287469288,
-      "eval_overall_precision": 0.9242053789731052,
-      "eval_overall_recall": 0.9333333333333333,
-      "eval_runtime": 0.5998,
-      "eval_samples_per_second": 283.424,
-      "eval_steps_per_second": 5.002,
+      "eval_loss": 0.048507072031497955,
+      "eval_overall_accuracy": 0.9828776581054958,
+      "eval_overall_f1": 0.9108669108669109,
+      "eval_overall_precision": 0.9009661835748792,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2721,
+      "eval_samples_per_second": 624.82,
+      "eval_steps_per_second": 11.026,
       "step": 3936
     },
     {
       "epoch": 42.0,
-      "grad_norm": 0.33734437823295593,
+      "grad_norm": 0.5790802240371704,
       "learning_rate": 2.9e-05,
-      "loss": 0.0576,
+      "loss": 0.0581,
       "step": 4032
     },
     {
       "epoch": 42.0,
-      "eval_LOCATION_f1": 0.9064039408866995,
+      "eval_LOCATION_f1": 0.8761904761904761,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.92,
+      "eval_LOCATION_precision": 0.8598130841121495,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.9111747851002865,
+      "eval_ORGANIZATION_f1": 0.881159420289855,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8932584269662921,
-      "eval_ORGANIZATION_recall": 0.9298245614035088,
+      "eval_ORGANIZATION_precision": 0.8735632183908046,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04518849030137062,
-      "eval_overall_accuracy": 0.9870201601767468,
-      "eval_overall_f1": 0.9276073619631902,
-      "eval_overall_precision": 0.9219512195121952,
-      "eval_overall_recall": 0.9333333333333333,
-      "eval_runtime": 0.5992,
-      "eval_samples_per_second": 283.689,
-      "eval_steps_per_second": 5.006,
+      "eval_loss": 0.04726165160536766,
+      "eval_overall_accuracy": 0.9828776581054958,
+      "eval_overall_f1": 0.9070904645476773,
+      "eval_overall_precision": 0.8983050847457628,
+      "eval_overall_recall": 0.9160493827160494,
+      "eval_runtime": 0.2765,
+      "eval_samples_per_second": 614.808,
+      "eval_steps_per_second": 10.85,
       "step": 4032
     },
     {
       "epoch": 43.0,
-      "grad_norm": 0.9250084757804871,
+      "grad_norm": 0.46468445658683777,
       "learning_rate": 2.8499999999999998e-05,
-      "loss": 0.057,
+      "loss": 0.0597,
       "step": 4128
     },
     {
@@ -1323,502 +1323,502 @@
       "eval_LOCATION_number": 103,
       "eval_LOCATION_precision": 0.8679245283018868,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8941176470588235,
+      "eval_ORGANIZATION_f1": 0.881159420289855,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8994082840236687,
+      "eval_ORGANIZATION_precision": 0.8735632183908046,
       "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04571167007088661,
-      "eval_overall_accuracy": 0.9853631593482464,
-      "eval_overall_f1": 0.9137931034482758,
-      "eval_overall_precision": 0.9115479115479116,
+      "eval_loss": 0.047872934490442276,
+      "eval_overall_accuracy": 0.9826014913007457,
+      "eval_overall_f1": 0.9093137254901962,
+      "eval_overall_precision": 0.902676399026764,
       "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.5947,
-      "eval_samples_per_second": 285.867,
-      "eval_steps_per_second": 5.045,
+      "eval_runtime": 0.2722,
+      "eval_samples_per_second": 624.585,
+      "eval_steps_per_second": 11.022,
       "step": 4128
     },
     {
       "epoch": 44.0,
-      "grad_norm": 1.0702989101409912,
+      "grad_norm": 0.5926026105880737,
       "learning_rate": 2.8000000000000003e-05,
-      "loss": 0.056,
+      "loss": 0.0568,
       "step": 4224
     },
     {
       "epoch": 44.0,
-      "eval_LOCATION_f1": 0.8932038834951457,
+      "eval_LOCATION_f1": 0.8720379146919431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8932038834951457,
+      "eval_LOCATION_precision": 0.8518518518518519,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8979591836734693,
+      "eval_ORGANIZATION_f1": 0.8875739644970414,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8953488372093024,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_ORGANIZATION_precision": 0.8982035928143712,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9657794676806084,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9621212121212122,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04532807692885399,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.91871921182266,
-      "eval_overall_precision": 0.9164619164619164,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.6003,
-      "eval_samples_per_second": 283.17,
-      "eval_steps_per_second": 4.997,
+      "eval_loss": 0.04811384528875351,
+      "eval_overall_accuracy": 0.9826014913007457,
+      "eval_overall_f1": 0.9088669950738916,
+      "eval_overall_precision": 0.9066339066339066,
+      "eval_overall_recall": 0.9111111111111111,
+      "eval_runtime": 0.2715,
+      "eval_samples_per_second": 626.105,
+      "eval_steps_per_second": 11.049,
       "step": 4224
     },
     {
       "epoch": 45.0,
-      "grad_norm": 1.9211422204971313,
+      "grad_norm": 1.227881669998169,
       "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.0557,
+      "loss": 0.0561,
       "step": 4320
     },
     {
       "epoch": 45.0,
-      "eval_LOCATION_f1": 0.8761904761904761,
+      "eval_LOCATION_f1": 0.8720379146919431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8598130841121495,
+      "eval_LOCATION_precision": 0.8518518518518519,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8914956011730205,
+      "eval_ORGANIZATION_f1": 0.8823529411764705,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8941176470588236,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8875739644970414,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.0455101914703846,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.9115479115479117,
-      "eval_overall_precision": 0.9070904645476773,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.6588,
-      "eval_samples_per_second": 258.043,
-      "eval_steps_per_second": 4.554,
+      "eval_loss": 0.04695962369441986,
+      "eval_overall_accuracy": 0.9834299917149959,
+      "eval_overall_f1": 0.9077490774907748,
+      "eval_overall_precision": 0.9044117647058824,
+      "eval_overall_recall": 0.9111111111111111,
+      "eval_runtime": 0.2758,
+      "eval_samples_per_second": 616.422,
+      "eval_steps_per_second": 10.878,
       "step": 4320
     },
     {
       "epoch": 46.0,
-      "grad_norm": 0.976703941822052,
+      "grad_norm": 1.023648738861084,
       "learning_rate": 2.7000000000000002e-05,
-      "loss": 0.0565,
+      "loss": 0.0552,
       "step": 4416
     },
     {
       "epoch": 46.0,
-      "eval_LOCATION_f1": 0.8651162790697675,
+      "eval_LOCATION_f1": 0.8720379146919431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8303571428571429,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8776119402985074,
+      "eval_LOCATION_precision": 0.8518518518518519,
+      "eval_LOCATION_recall": 0.8932038834951457,
+      "eval_ORGANIZATION_f1": 0.890207715133531,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8963414634146342,
-      "eval_ORGANIZATION_recall": 0.8596491228070176,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.9036144578313253,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04744512960314751,
-      "eval_overall_accuracy": 0.9828776581054958,
-      "eval_overall_f1": 0.902829028290283,
-      "eval_overall_precision": 0.8995098039215687,
-      "eval_overall_recall": 0.9061728395061729,
-      "eval_runtime": 0.5904,
-      "eval_samples_per_second": 287.926,
-      "eval_steps_per_second": 5.081,
+      "eval_loss": 0.04779404029250145,
+      "eval_overall_accuracy": 0.983706158519746,
+      "eval_overall_f1": 0.9111111111111111,
+      "eval_overall_precision": 0.9111111111111111,
+      "eval_overall_recall": 0.9111111111111111,
+      "eval_runtime": 0.2711,
+      "eval_samples_per_second": 627.02,
+      "eval_steps_per_second": 11.065,
       "step": 4416
     },
     {
       "epoch": 47.0,
-      "grad_norm": 0.8773436546325684,
+      "grad_norm": 0.36909419298171997,
       "learning_rate": 2.6500000000000004e-05,
-      "loss": 0.0534,
+      "loss": 0.0562,
       "step": 4512
     },
     {
       "epoch": 47.0,
-      "eval_LOCATION_f1": 0.8932038834951457,
+      "eval_LOCATION_f1": 0.8846153846153847,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8932038834951457,
+      "eval_LOCATION_precision": 0.8761904761904762,
       "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8985507246376812,
+      "eval_ORGANIZATION_f1": 0.8793103448275862,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8908045977011494,
-      "eval_ORGANIZATION_recall": 0.9064327485380117,
+      "eval_ORGANIZATION_precision": 0.864406779661017,
+      "eval_ORGANIZATION_recall": 0.8947368421052632,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04476306214928627,
-      "eval_overall_accuracy": 0.9859154929577465,
-      "eval_overall_f1": 0.920049200492005,
-      "eval_overall_precision": 0.9166666666666666,
-      "eval_overall_recall": 0.9234567901234568,
-      "eval_runtime": 0.5945,
-      "eval_samples_per_second": 285.942,
-      "eval_steps_per_second": 5.046,
+      "eval_loss": 0.046062689274549484,
+      "eval_overall_accuracy": 0.983982325324496,
+      "eval_overall_f1": 0.9095354523227382,
+      "eval_overall_precision": 0.9007263922518159,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.272,
+      "eval_samples_per_second": 624.965,
+      "eval_steps_per_second": 11.029,
       "step": 4512
     },
     {
       "epoch": 48.0,
-      "grad_norm": 1.278594970703125,
+      "grad_norm": 0.6459914445877075,
       "learning_rate": 2.6000000000000002e-05,
-      "loss": 0.0523,
+      "loss": 0.0533,
       "step": 4608
     },
     {
       "epoch": 48.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.8826291079812206,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8961424332344213,
+      "eval_LOCATION_precision": 0.8545454545454545,
+      "eval_LOCATION_recall": 0.912621359223301,
+      "eval_ORGANIZATION_f1": 0.8895522388059701,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9096385542168675,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.9085365853658537,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04515359550714493,
-      "eval_overall_accuracy": 0.9859154929577465,
-      "eval_overall_f1": 0.9149198520345252,
-      "eval_overall_precision": 0.9137931034482759,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.5953,
-      "eval_samples_per_second": 285.59,
-      "eval_steps_per_second": 5.04,
+      "eval_loss": 0.04743105545639992,
+      "eval_overall_accuracy": 0.983706158519746,
+      "eval_overall_f1": 0.9135802469135802,
+      "eval_overall_precision": 0.9135802469135802,
+      "eval_overall_recall": 0.9135802469135802,
+      "eval_runtime": 0.2736,
+      "eval_samples_per_second": 621.325,
+      "eval_steps_per_second": 10.965,
       "step": 4608
     },
     {
       "epoch": 49.0,
-      "grad_norm": 0.56380695104599,
+      "grad_norm": 0.17258469760417938,
       "learning_rate": 2.5500000000000003e-05,
-      "loss": 0.0523,
+      "loss": 0.0522,
       "step": 4704
     },
     {
       "epoch": 49.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.8785046728971964,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8856304985337242,
+      "eval_LOCATION_precision": 0.8468468468468469,
+      "eval_LOCATION_recall": 0.912621359223301,
+      "eval_ORGANIZATION_f1": 0.8771929824561403,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.888235294117647,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8771929824561403,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04556996375322342,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9104294478527607,
-      "eval_overall_precision": 0.9048780487804878,
+      "eval_loss": 0.04609883204102516,
+      "eval_overall_accuracy": 0.9842584921292461,
+      "eval_overall_f1": 0.9070904645476773,
+      "eval_overall_precision": 0.8983050847457628,
       "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.5918,
-      "eval_samples_per_second": 287.26,
-      "eval_steps_per_second": 5.069,
+      "eval_runtime": 0.2713,
+      "eval_samples_per_second": 626.665,
+      "eval_steps_per_second": 11.059,
       "step": 4704
     },
     {
       "epoch": 50.0,
-      "grad_norm": 1.6369190216064453,
+      "grad_norm": 1.7320947647094727,
       "learning_rate": 2.5e-05,
-      "loss": 0.0509,
+      "loss": 0.052,
       "step": 4800
     },
     {
       "epoch": 50.0,
-      "eval_LOCATION_f1": 0.8773584905660378,
+      "eval_LOCATION_f1": 0.8878504672897196,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8532110091743119,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8802395209580838,
+      "eval_LOCATION_precision": 0.8558558558558559,
+      "eval_LOCATION_recall": 0.9223300970873787,
+      "eval_ORGANIZATION_f1": 0.8869565217391304,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.901840490797546,
-      "eval_ORGANIZATION_recall": 0.8596491228070176,
+      "eval_ORGANIZATION_precision": 0.8793103448275862,
+      "eval_ORGANIZATION_recall": 0.8947368421052632,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.046390656381845474,
-      "eval_overall_accuracy": 0.9828776581054958,
-      "eval_overall_f1": 0.9084158415841586,
-      "eval_overall_precision": 0.9106699751861043,
-      "eval_overall_recall": 0.9061728395061729,
-      "eval_runtime": 0.6054,
-      "eval_samples_per_second": 280.8,
-      "eval_steps_per_second": 4.955,
+      "eval_loss": 0.04637792333960533,
+      "eval_overall_accuracy": 0.983982325324496,
+      "eval_overall_f1": 0.9135200974421438,
+      "eval_overall_precision": 0.9014423076923077,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.271,
+      "eval_samples_per_second": 627.338,
+      "eval_steps_per_second": 11.071,
       "step": 4800
     },
     {
       "epoch": 51.0,
-      "grad_norm": 0.7033815383911133,
+      "grad_norm": 0.8508228063583374,
       "learning_rate": 2.45e-05,
-      "loss": 0.052,
+      "loss": 0.054,
       "step": 4896
     },
     {
       "epoch": 51.0,
-      "eval_LOCATION_f1": 0.8663594470046084,
+      "eval_LOCATION_f1": 0.8930232558139534,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8245614035087719,
-      "eval_LOCATION_recall": 0.912621359223301,
-      "eval_ORGANIZATION_f1": 0.8795180722891567,
+      "eval_LOCATION_precision": 0.8571428571428571,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8869047619047619,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.906832298136646,
-      "eval_ORGANIZATION_recall": 0.8538011695906432,
+      "eval_ORGANIZATION_precision": 0.9030303030303031,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04676185920834541,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.9050554870530209,
-      "eval_overall_precision": 0.9039408866995073,
-      "eval_overall_recall": 0.9061728395061729,
-      "eval_runtime": 0.5936,
-      "eval_samples_per_second": 286.386,
-      "eval_steps_per_second": 5.054,
+      "eval_loss": 0.046650972217321396,
+      "eval_overall_accuracy": 0.9853631593482464,
+      "eval_overall_f1": 0.915129151291513,
+      "eval_overall_precision": 0.9117647058823529,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2718,
+      "eval_samples_per_second": 625.424,
+      "eval_steps_per_second": 11.037,
       "step": 4896
     },
     {
       "epoch": 52.0,
-      "grad_norm": 0.6778050065040588,
+      "grad_norm": 0.6740456223487854,
       "learning_rate": 2.4e-05,
-      "loss": 0.0511,
+      "loss": 0.0525,
       "step": 4992
     },
     {
       "epoch": 52.0,
-      "eval_LOCATION_f1": 0.8975609756097561,
+      "eval_LOCATION_f1": 0.9158878504672898,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9019607843137255,
-      "eval_LOCATION_recall": 0.8932038834951457,
-      "eval_ORGANIZATION_f1": 0.8979591836734693,
+      "eval_LOCATION_precision": 0.8828828828828829,
+      "eval_LOCATION_recall": 0.9514563106796117,
+      "eval_ORGANIZATION_f1": 0.872093023255814,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8953488372093024,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_ORGANIZATION_precision": 0.8670520231213873,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04431942105293274,
-      "eval_overall_accuracy": 0.9861916597624966,
-      "eval_overall_f1": 0.9209876543209876,
-      "eval_overall_precision": 0.9209876543209876,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.6009,
-      "eval_samples_per_second": 282.922,
-      "eval_steps_per_second": 4.993,
+      "eval_loss": 0.04597054049372673,
+      "eval_overall_accuracy": 0.983982325324496,
+      "eval_overall_f1": 0.9146341463414634,
+      "eval_overall_precision": 0.9036144578313253,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2723,
+      "eval_samples_per_second": 624.335,
+      "eval_steps_per_second": 11.018,
       "step": 4992
     },
     {
       "epoch": 53.0,
-      "grad_norm": 0.7755054831504822,
+      "grad_norm": 1.4370176792144775,
       "learning_rate": 2.35e-05,
-      "loss": 0.051,
+      "loss": 0.0501,
       "step": 5088
     },
     {
       "epoch": 53.0,
-      "eval_LOCATION_f1": 0.8773584905660378,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8532110091743119,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8875739644970414,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8816568047337278,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8982035928143712,
-      "eval_ORGANIZATION_recall": 0.8771929824561403,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8922155688622755,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045676685869693756,
-      "eval_overall_accuracy": 0.983982325324496,
-      "eval_overall_f1": 0.9102091020910209,
-      "eval_overall_precision": 0.9068627450980392,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.61,
-      "eval_samples_per_second": 278.668,
-      "eval_steps_per_second": 4.918,
+      "eval_loss": 0.04658225178718567,
+      "eval_overall_accuracy": 0.9848108257387462,
+      "eval_overall_f1": 0.9164619164619164,
+      "eval_overall_precision": 0.9119804400977995,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2733,
+      "eval_samples_per_second": 622.121,
+      "eval_steps_per_second": 10.979,
       "step": 5088
     },
     {
       "epoch": 54.0,
-      "grad_norm": 0.47155851125717163,
+      "grad_norm": 1.064245343208313,
       "learning_rate": 2.3000000000000003e-05,
-      "loss": 0.0483,
+      "loss": 0.0498,
       "step": 5184
     },
     {
       "epoch": 54.0,
-      "eval_LOCATION_f1": 0.9029126213592235,
+      "eval_LOCATION_f1": 0.8962264150943398,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.9029126213592233,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.9017341040462428,
+      "eval_LOCATION_precision": 0.8715596330275229,
+      "eval_LOCATION_recall": 0.9223300970873787,
+      "eval_ORGANIZATION_f1": 0.880466472303207,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8914285714285715,
-      "eval_ORGANIZATION_recall": 0.9122807017543859,
+      "eval_ORGANIZATION_precision": 0.877906976744186,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.044774431735277176,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9238329238329238,
-      "eval_overall_precision": 0.9193154034229829,
-      "eval_overall_recall": 0.928395061728395,
-      "eval_runtime": 0.5995,
-      "eval_samples_per_second": 283.593,
-      "eval_steps_per_second": 5.005,
+      "eval_loss": 0.04490743577480316,
+      "eval_overall_accuracy": 0.983982325324496,
+      "eval_overall_f1": 0.9130966952264382,
+      "eval_overall_precision": 0.9053398058252428,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2734,
+      "eval_samples_per_second": 621.817,
+      "eval_steps_per_second": 10.973,
       "step": 5184
     },
     {
       "epoch": 55.0,
-      "grad_norm": 1.5237046480178833,
+      "grad_norm": 2.0687813758850098,
       "learning_rate": 2.25e-05,
-      "loss": 0.0525,
+      "loss": 0.0504,
       "step": 5280
     },
     {
       "epoch": 55.0,
-      "eval_LOCATION_f1": 0.8899521531100479,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8773584905660378,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8953488372093024,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8875739644970414,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8901734104046243,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_ORGANIZATION_precision": 0.8982035928143712,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04513506591320038,
+      "eval_loss": 0.045612651854753494,
       "eval_overall_accuracy": 0.9853631593482464,
-      "eval_overall_f1": 0.9177914110429448,
-      "eval_overall_precision": 0.9121951219512195,
+      "eval_overall_f1": 0.918918918918919,
+      "eval_overall_precision": 0.9144254278728606,
       "eval_overall_recall": 0.9234567901234568,
-      "eval_runtime": 0.6045,
-      "eval_samples_per_second": 281.204,
-      "eval_steps_per_second": 4.962,
+      "eval_runtime": 0.2766,
+      "eval_samples_per_second": 614.623,
+      "eval_steps_per_second": 10.846,
       "step": 5280
     },
     {
       "epoch": 56.0,
-      "grad_norm": 0.6528781652450562,
+      "grad_norm": 1.0828304290771484,
       "learning_rate": 2.2000000000000003e-05,
-      "loss": 0.0479,
+      "loss": 0.0486,
       "step": 5376
     },
     {
       "epoch": 56.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8914956011730205,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.881159420289855,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8941176470588236,
+      "eval_ORGANIZATION_precision": 0.8735632183908046,
       "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04449770227074623,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.915129151291513,
-      "eval_overall_precision": 0.9117647058823529,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.5981,
-      "eval_samples_per_second": 284.219,
-      "eval_steps_per_second": 5.016,
+      "eval_loss": 0.045335978269577026,
+      "eval_overall_accuracy": 0.9845346589339962,
+      "eval_overall_f1": 0.9146341463414634,
+      "eval_overall_precision": 0.9036144578313253,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2731,
+      "eval_samples_per_second": 622.495,
+      "eval_steps_per_second": 10.985,
       "step": 5376
     },
     {
       "epoch": 57.0,
-      "grad_norm": 1.4258739948272705,
+      "grad_norm": 1.297290563583374,
       "learning_rate": 2.15e-05,
-      "loss": 0.0486,
+      "loss": 0.0497,
       "step": 5472
     },
     {
       "epoch": 57.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.8940092165898617,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8888888888888888,
+      "eval_LOCATION_precision": 0.8508771929824561,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8875739644970414,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8888888888888888,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
-      "eval_PERSON_f1": 0.9657794676806084,
+      "eval_ORGANIZATION_precision": 0.8982035928143712,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
-      "eval_PERSON_precision": 0.9621212121212122,
+      "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.044527389109134674,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9128834355828221,
-      "eval_overall_precision": 0.9073170731707317,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.597,
-      "eval_samples_per_second": 284.74,
-      "eval_steps_per_second": 5.025,
+      "eval_loss": 0.045681606978178024,
+      "eval_overall_accuracy": 0.9845346589339962,
+      "eval_overall_f1": 0.9155446756425949,
+      "eval_overall_precision": 0.9077669902912622,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2721,
+      "eval_samples_per_second": 624.822,
+      "eval_steps_per_second": 11.026,
       "step": 5472
     },
     {
       "epoch": 58.0,
-      "grad_norm": 0.6324487924575806,
+      "grad_norm": 0.472746878862381,
       "learning_rate": 2.1e-05,
-      "loss": 0.0457,
+      "loss": 0.0487,
       "step": 5568
     },
     {
       "epoch": 58.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8947368421052632,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.890207715133531,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8947368421052632,
-      "eval_ORGANIZATION_recall": 0.8947368421052632,
+      "eval_ORGANIZATION_precision": 0.9036144578313253,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04367503523826599,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9164619164619164,
-      "eval_overall_precision": 0.9119804400977995,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.6033,
-      "eval_samples_per_second": 281.778,
-      "eval_steps_per_second": 4.973,
+      "eval_loss": 0.04600377753376961,
+      "eval_overall_accuracy": 0.9853631593482464,
+      "eval_overall_f1": 0.920049200492005,
+      "eval_overall_precision": 0.9166666666666666,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2747,
+      "eval_samples_per_second": 618.786,
+      "eval_steps_per_second": 10.92,
       "step": 5568
     },
     {
       "epoch": 59.0,
-      "grad_norm": 1.5328294038772583,
+      "grad_norm": 0.6368147730827332,
       "learning_rate": 2.05e-05,
-      "loss": 0.0478,
+      "loss": 0.0473,
       "step": 5664
     },
     {
       "epoch": 59.0,
-      "eval_LOCATION_f1": 0.8773584905660378,
+      "eval_LOCATION_f1": 0.8878504672897196,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8532110091743119,
-      "eval_LOCATION_recall": 0.9029126213592233,
+      "eval_LOCATION_precision": 0.8558558558558559,
+      "eval_LOCATION_recall": 0.9223300970873787,
       "eval_ORGANIZATION_f1": 0.8816568047337278,
       "eval_ORGANIZATION_number": 171,
       "eval_ORGANIZATION_precision": 0.8922155688622755,
@@ -1827,1295 +1827,1295 @@
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04407254979014397,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9088669950738916,
-      "eval_overall_precision": 0.9066339066339066,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6034,
-      "eval_samples_per_second": 281.724,
-      "eval_steps_per_second": 4.972,
+      "eval_loss": 0.045563556253910065,
+      "eval_overall_accuracy": 0.9850869925434963,
+      "eval_overall_f1": 0.9115479115479117,
+      "eval_overall_precision": 0.9070904645476773,
+      "eval_overall_recall": 0.9160493827160494,
+      "eval_runtime": 0.2709,
+      "eval_samples_per_second": 627.644,
+      "eval_steps_per_second": 11.076,
       "step": 5664
     },
     {
       "epoch": 60.0,
-      "grad_norm": 0.5337828993797302,
+      "grad_norm": 0.35717150568962097,
       "learning_rate": 2e-05,
-      "loss": 0.0472,
+      "loss": 0.0463,
       "step": 5760
     },
     {
       "epoch": 60.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.8878504672897196,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8849557522123894,
+      "eval_LOCATION_precision": 0.8558558558558559,
+      "eval_LOCATION_recall": 0.9223300970873787,
+      "eval_ORGANIZATION_f1": 0.8809523809523809,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8928571428571429,
-      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_ORGANIZATION_precision": 0.896969696969697,
+      "eval_ORGANIZATION_recall": 0.8654970760233918,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04403013363480568,
-      "eval_overall_accuracy": 0.9850869925434963,
+      "eval_loss": 0.04541854187846184,
+      "eval_overall_accuracy": 0.9853631593482464,
       "eval_overall_f1": 0.9113300492610836,
       "eval_overall_precision": 0.9090909090909091,
       "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.5975,
-      "eval_samples_per_second": 284.5,
-      "eval_steps_per_second": 5.021,
+      "eval_runtime": 0.2737,
+      "eval_samples_per_second": 621.148,
+      "eval_steps_per_second": 10.961,
       "step": 5760
     },
     {
       "epoch": 61.0,
-      "grad_norm": 0.9366702437400818,
+      "grad_norm": 1.0991231203079224,
       "learning_rate": 1.9500000000000003e-05,
-      "loss": 0.049,
+      "loss": 0.0486,
       "step": 5856
     },
     {
       "epoch": 61.0,
-      "eval_LOCATION_f1": 0.8942307692307693,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8857142857142857,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8895348837209303,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8809523809523809,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.884393063583815,
-      "eval_ORGANIZATION_recall": 0.8947368421052632,
+      "eval_ORGANIZATION_precision": 0.896969696969697,
+      "eval_ORGANIZATION_recall": 0.8654970760233918,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04451928660273552,
-      "eval_overall_accuracy": 0.9853631593482464,
-      "eval_overall_f1": 0.9164619164619164,
-      "eval_overall_precision": 0.9119804400977995,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.6062,
-      "eval_samples_per_second": 280.441,
-      "eval_steps_per_second": 4.949,
+      "eval_loss": 0.045604679733514786,
+      "eval_overall_accuracy": 0.9856393261529964,
+      "eval_overall_f1": 0.916256157635468,
+      "eval_overall_precision": 0.914004914004914,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2735,
+      "eval_samples_per_second": 621.534,
+      "eval_steps_per_second": 10.968,
       "step": 5856
     },
     {
       "epoch": 62.0,
-      "grad_norm": 1.8338637351989746,
+      "grad_norm": 1.2870984077453613,
       "learning_rate": 1.9e-05,
-      "loss": 0.0476,
+      "loss": 0.0484,
       "step": 5952
     },
     {
       "epoch": 62.0,
-      "eval_LOCATION_f1": 0.869158878504673,
+      "eval_LOCATION_f1": 0.8930232558139534,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8378378378378378,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.875739644970414,
+      "eval_LOCATION_precision": 0.8571428571428571,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8809523809523809,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8862275449101796,
+      "eval_ORGANIZATION_precision": 0.896969696969697,
       "eval_ORGANIZATION_recall": 0.8654970760233918,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045562200248241425,
-      "eval_overall_accuracy": 0.983982325324496,
-      "eval_overall_f1": 0.9041769041769042,
-      "eval_overall_precision": 0.8997555012224939,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6542,
-      "eval_samples_per_second": 259.871,
-      "eval_steps_per_second": 4.586,
+      "eval_loss": 0.04645540565252304,
+      "eval_overall_accuracy": 0.9850869925434963,
+      "eval_overall_f1": 0.9126691266912669,
+      "eval_overall_precision": 0.9093137254901961,
+      "eval_overall_recall": 0.9160493827160494,
+      "eval_runtime": 0.2716,
+      "eval_samples_per_second": 626.017,
+      "eval_steps_per_second": 11.047,
       "step": 5952
     },
     {
       "epoch": 63.0,
-      "grad_norm": 2.0841386318206787,
+      "grad_norm": 1.264487862586975,
       "learning_rate": 1.85e-05,
-      "loss": 0.0457,
+      "loss": 0.0461,
       "step": 6048
     },
     {
       "epoch": 63.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.9116279069767441,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8979591836734693,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9514563106796117,
+      "eval_ORGANIZATION_f1": 0.8908554572271385,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8953488372093024,
-      "eval_ORGANIZATION_recall": 0.9005847953216374,
+      "eval_ORGANIZATION_precision": 0.8988095238095238,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.043963029980659485,
-      "eval_overall_accuracy": 0.9859154929577465,
-      "eval_overall_f1": 0.9177914110429448,
-      "eval_overall_precision": 0.9121951219512195,
-      "eval_overall_recall": 0.9234567901234568,
-      "eval_runtime": 0.6065,
-      "eval_samples_per_second": 280.291,
-      "eval_steps_per_second": 4.946,
+      "eval_loss": 0.04506193846464157,
+      "eval_overall_accuracy": 0.9856393261529964,
+      "eval_overall_f1": 0.9215686274509803,
+      "eval_overall_precision": 0.9148418491484185,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2722,
+      "eval_samples_per_second": 624.428,
+      "eval_steps_per_second": 11.019,
       "step": 6048
     },
     {
       "epoch": 64.0,
-      "grad_norm": 0.9337228536605835,
+      "grad_norm": 0.35966232419013977,
       "learning_rate": 1.8e-05,
-      "loss": 0.0446,
+      "loss": 0.0455,
       "step": 6144
     },
     {
       "epoch": 64.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8882352941176469,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8842729970326408,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.893491124260355,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
+      "eval_ORGANIZATION_precision": 0.8975903614457831,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04478294402360916,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9137931034482758,
-      "eval_overall_precision": 0.9115479115479116,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.5976,
-      "eval_samples_per_second": 284.482,
-      "eval_steps_per_second": 5.02,
+      "eval_loss": 0.04505002498626709,
+      "eval_overall_accuracy": 0.9853631593482464,
+      "eval_overall_f1": 0.916256157635468,
+      "eval_overall_precision": 0.914004914004914,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2717,
+      "eval_samples_per_second": 625.717,
+      "eval_steps_per_second": 11.042,
       "step": 6144
     },
     {
       "epoch": 65.0,
-      "grad_norm": 1.0935461521148682,
+      "grad_norm": 0.5010517239570618,
       "learning_rate": 1.75e-05,
-      "loss": 0.0465,
+      "loss": 0.0472,
       "step": 6240
     },
     {
       "epoch": 65.0,
-      "eval_LOCATION_f1": 0.8826291079812206,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8545454545454545,
-      "eval_LOCATION_recall": 0.912621359223301,
-      "eval_ORGANIZATION_f1": 0.8908554572271385,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8869047619047619,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8988095238095238,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
+      "eval_ORGANIZATION_precision": 0.9030303030303031,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04469531029462814,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.914004914004914,
-      "eval_overall_precision": 0.9095354523227384,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.603,
-      "eval_samples_per_second": 281.935,
-      "eval_steps_per_second": 4.975,
+      "eval_loss": 0.04529740288853645,
+      "eval_overall_accuracy": 0.9859154929577465,
+      "eval_overall_f1": 0.91871921182266,
+      "eval_overall_precision": 0.9164619164619164,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 623.611,
+      "eval_steps_per_second": 11.005,
       "step": 6240
     },
     {
       "epoch": 66.0,
-      "grad_norm": 0.588843584060669,
+      "grad_norm": 0.859156608581543,
       "learning_rate": 1.7000000000000003e-05,
-      "loss": 0.0456,
+      "loss": 0.0453,
       "step": 6336
     },
     {
       "epoch": 66.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8842729970326408,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8975903614457831,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04511041194200516,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9122373300370827,
-      "eval_overall_precision": 0.9133663366336634,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5959,
-      "eval_samples_per_second": 285.269,
-      "eval_steps_per_second": 5.034,
+      "eval_loss": 0.04506004601716995,
+      "eval_overall_accuracy": 0.9864678265672466,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.273,
+      "eval_samples_per_second": 622.688,
+      "eval_steps_per_second": 10.989,
       "step": 6336
     },
     {
       "epoch": 67.0,
-      "grad_norm": 0.40730783343315125,
+      "grad_norm": 0.3856222927570343,
       "learning_rate": 1.65e-05,
-      "loss": 0.043,
+      "loss": 0.045,
       "step": 6432
     },
     {
       "epoch": 67.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.874251497005988,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8895522388059701,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8957055214723927,
-      "eval_ORGANIZATION_recall": 0.8538011695906432,
+      "eval_ORGANIZATION_precision": 0.9085365853658537,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04662081599235535,
-      "eval_overall_accuracy": 0.9823253244959956,
-      "eval_overall_f1": 0.9070631970260223,
-      "eval_overall_precision": 0.9104477611940298,
-      "eval_overall_recall": 0.9037037037037037,
-      "eval_runtime": 0.5968,
-      "eval_samples_per_second": 284.867,
-      "eval_steps_per_second": 5.027,
+      "eval_loss": 0.04503821209073067,
+      "eval_overall_accuracy": 0.9856393261529964,
+      "eval_overall_f1": 0.9198520345252774,
+      "eval_overall_precision": 0.9187192118226601,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2725,
+      "eval_samples_per_second": 623.806,
+      "eval_steps_per_second": 11.008,
       "step": 6432
     },
     {
       "epoch": 68.0,
-      "grad_norm": 1.010898470878601,
+      "grad_norm": 0.5940878987312317,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.0441,
+      "loss": 0.0466,
       "step": 6528
     },
     {
       "epoch": 68.0,
-      "eval_LOCATION_f1": 0.8909952606635072,
+      "eval_LOCATION_f1": 0.9201877934272301,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8703703703703703,
-      "eval_LOCATION_recall": 0.912621359223301,
-      "eval_ORGANIZATION_f1": 0.879765395894428,
+      "eval_LOCATION_precision": 0.8909090909090909,
+      "eval_LOCATION_recall": 0.9514563106796117,
+      "eval_ORGANIZATION_f1": 0.890207715133531,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8823529411764706,
+      "eval_ORGANIZATION_precision": 0.9036144578313253,
       "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04551886394619942,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9115479115479117,
-      "eval_overall_precision": 0.9070904645476773,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.6631,
-      "eval_samples_per_second": 256.37,
-      "eval_steps_per_second": 4.524,
+      "eval_loss": 0.044038835912942886,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9236453201970444,
+      "eval_overall_precision": 0.9213759213759214,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2728,
+      "eval_samples_per_second": 623.229,
+      "eval_steps_per_second": 10.998,
       "step": 6528
     },
     {
       "epoch": 69.0,
-      "grad_norm": 0.356332927942276,
+      "grad_norm": 1.601568579673767,
       "learning_rate": 1.55e-05,
-      "loss": 0.0439,
+      "loss": 0.046,
       "step": 6624
     },
     {
       "epoch": 69.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.8971962616822431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.879765395894428,
+      "eval_LOCATION_precision": 0.8648648648648649,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8895522388059701,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8823529411764706,
-      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_ORGANIZATION_precision": 0.9085365853658537,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04542705416679382,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9102091020910209,
-      "eval_overall_precision": 0.9068627450980392,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.5976,
-      "eval_samples_per_second": 284.455,
-      "eval_steps_per_second": 5.02,
+      "eval_loss": 0.04461726173758507,
+      "eval_overall_accuracy": 0.9853631593482464,
+      "eval_overall_f1": 0.9173859432799014,
+      "eval_overall_precision": 0.916256157635468,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2738,
+      "eval_samples_per_second": 620.821,
+      "eval_steps_per_second": 10.956,
       "step": 6624
     },
     {
       "epoch": 70.0,
-      "grad_norm": 0.5880185961723328,
+      "grad_norm": 2.521439552307129,
       "learning_rate": 1.5e-05,
-      "loss": 0.0439,
+      "loss": 0.0436,
       "step": 6720
     },
     {
       "epoch": 70.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.8971962616822431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8882352941176469,
+      "eval_LOCATION_precision": 0.8648648648648649,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8928571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.893491124260355,
-      "eval_ORGANIZATION_recall": 0.8830409356725146,
+      "eval_ORGANIZATION_precision": 0.9090909090909091,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.044994134455919266,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.9126691266912669,
-      "eval_overall_precision": 0.9093137254901961,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.597,
-      "eval_samples_per_second": 284.742,
-      "eval_steps_per_second": 5.025,
+      "eval_loss": 0.04403162747621536,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.91871921182266,
+      "eval_overall_precision": 0.9164619164619164,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.275,
+      "eval_samples_per_second": 618.275,
+      "eval_steps_per_second": 10.911,
       "step": 6720
     },
     {
       "epoch": 71.0,
-      "grad_norm": 0.9274646639823914,
+      "grad_norm": 0.8414034247398376,
       "learning_rate": 1.45e-05,
-      "loss": 0.0428,
+      "loss": 0.0445,
       "step": 6816
     },
     {
       "epoch": 71.0,
-      "eval_LOCATION_f1": 0.8952380952380954,
+      "eval_LOCATION_f1": 0.9074074074074074,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8785046728971962,
-      "eval_LOCATION_recall": 0.912621359223301,
-      "eval_ORGANIZATION_f1": 0.8888888888888888,
+      "eval_LOCATION_precision": 0.8672566371681416,
+      "eval_LOCATION_recall": 0.9514563106796117,
+      "eval_ORGANIZATION_f1": 0.8908554572271385,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8888888888888888,
-      "eval_ORGANIZATION_recall": 0.8888888888888888,
+      "eval_ORGANIZATION_precision": 0.8988095238095238,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04472970589995384,
-      "eval_overall_accuracy": 0.9859154929577465,
-      "eval_overall_f1": 0.9164619164619164,
-      "eval_overall_precision": 0.9119804400977995,
-      "eval_overall_recall": 0.9209876543209876,
-      "eval_runtime": 0.5987,
-      "eval_samples_per_second": 283.931,
-      "eval_steps_per_second": 5.011,
+      "eval_loss": 0.0446244440972805,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.9204406364749083,
+      "eval_overall_precision": 0.912621359223301,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2729,
+      "eval_samples_per_second": 623.038,
+      "eval_steps_per_second": 10.995,
       "step": 6816
     },
     {
       "epoch": 72.0,
-      "grad_norm": 0.35504868626594543,
+      "grad_norm": 0.7483597993850708,
       "learning_rate": 1.4000000000000001e-05,
-      "loss": 0.0438,
+      "loss": 0.0437,
       "step": 6912
     },
     {
       "epoch": 72.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8795180722891567,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8835820895522388,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.906832298136646,
-      "eval_ORGANIZATION_recall": 0.8538011695906432,
+      "eval_ORGANIZATION_precision": 0.9024390243902439,
+      "eval_ORGANIZATION_recall": 0.8654970760233918,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.046420272439718246,
-      "eval_overall_accuracy": 0.9828776581054958,
-      "eval_overall_f1": 0.9093167701863354,
-      "eval_overall_precision": 0.915,
-      "eval_overall_recall": 0.9037037037037037,
-      "eval_runtime": 0.5929,
-      "eval_samples_per_second": 286.713,
-      "eval_steps_per_second": 5.06,
+      "eval_loss": 0.04594691842794418,
+      "eval_overall_accuracy": 0.9853631593482464,
+      "eval_overall_f1": 0.9160493827160494,
+      "eval_overall_precision": 0.9160493827160494,
+      "eval_overall_recall": 0.9160493827160494,
+      "eval_runtime": 0.2711,
+      "eval_samples_per_second": 627.021,
+      "eval_steps_per_second": 11.065,
       "step": 6912
     },
     {
       "epoch": 73.0,
-      "grad_norm": 1.1146135330200195,
+      "grad_norm": 1.4202793836593628,
       "learning_rate": 1.3500000000000001e-05,
-      "loss": 0.0431,
+      "loss": 0.0434,
       "step": 7008
     },
     {
       "epoch": 73.0,
-      "eval_LOCATION_f1": 0.8867924528301887,
+      "eval_LOCATION_f1": 0.9116279069767441,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8623853211009175,
-      "eval_LOCATION_recall": 0.912621359223301,
-      "eval_ORGANIZATION_f1": 0.8823529411764705,
+      "eval_LOCATION_precision": 0.875,
+      "eval_LOCATION_recall": 0.9514563106796117,
+      "eval_ORGANIZATION_f1": 0.8928571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8875739644970414,
+      "eval_ORGANIZATION_precision": 0.9090909090909091,
       "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.044833190739154816,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9115479115479117,
-      "eval_overall_precision": 0.9070904645476773,
-      "eval_overall_recall": 0.9160493827160494,
-      "eval_runtime": 0.606,
-      "eval_samples_per_second": 280.509,
-      "eval_steps_per_second": 4.95,
+      "eval_loss": 0.0443597137928009,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.922509225092251,
+      "eval_overall_precision": 0.9191176470588235,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2743,
+      "eval_samples_per_second": 619.819,
+      "eval_steps_per_second": 10.938,
       "step": 7008
     },
     {
       "epoch": 74.0,
-      "grad_norm": 0.3997597396373749,
+      "grad_norm": 0.6511960625648499,
       "learning_rate": 1.3000000000000001e-05,
-      "loss": 0.0415,
+      "loss": 0.0441,
       "step": 7104
     },
     {
       "epoch": 74.0,
-      "eval_LOCATION_f1": 0.8773584905660378,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8532110091743119,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.875739644970414,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8928571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8862275449101796,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9090909090909091,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04576844349503517,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6179,
-      "eval_samples_per_second": 275.116,
-      "eval_steps_per_second": 4.855,
+      "eval_loss": 0.044465843588113785,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.9211822660098522,
+      "eval_overall_precision": 0.918918918918919,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2712,
+      "eval_samples_per_second": 626.86,
+      "eval_steps_per_second": 11.062,
       "step": 7104
     },
     {
       "epoch": 75.0,
-      "grad_norm": 0.6900985836982727,
+      "grad_norm": 0.9006913900375366,
       "learning_rate": 1.25e-05,
-      "loss": 0.0429,
+      "loss": 0.0439,
       "step": 7200
     },
     {
       "epoch": 75.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8928571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9090909090909091,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.046131208539009094,
-      "eval_overall_accuracy": 0.983982325324496,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6007,
-      "eval_samples_per_second": 283.003,
-      "eval_steps_per_second": 4.994,
+      "eval_loss": 0.04460127279162407,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.9211822660098522,
+      "eval_overall_precision": 0.918918918918919,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2727,
+      "eval_samples_per_second": 623.458,
+      "eval_steps_per_second": 11.002,
       "step": 7200
     },
     {
       "epoch": 76.0,
-      "grad_norm": 0.35220813751220703,
+      "grad_norm": 0.4795643389225006,
       "learning_rate": 1.2e-05,
-      "loss": 0.0426,
+      "loss": 0.042,
       "step": 7296
     },
     {
       "epoch": 76.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9023255813953488,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8660714285714286,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8875739644970414,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.8982035928143712,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04544881358742714,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.5968,
-      "eval_samples_per_second": 284.861,
-      "eval_steps_per_second": 5.027,
+      "eval_loss": 0.044701043516397476,
+      "eval_overall_accuracy": 0.9859154929577465,
+      "eval_overall_f1": 0.9177914110429448,
+      "eval_overall_precision": 0.9121951219512195,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2713,
+      "eval_samples_per_second": 626.658,
+      "eval_steps_per_second": 11.059,
       "step": 7296
     },
     {
       "epoch": 77.0,
-      "grad_norm": 0.7620628476142883,
+      "grad_norm": 0.5350348949432373,
       "learning_rate": 1.1500000000000002e-05,
-      "loss": 0.043,
+      "loss": 0.0428,
       "step": 7392
     },
     {
       "epoch": 77.0,
-      "eval_LOCATION_f1": 0.8857142857142858,
+      "eval_LOCATION_f1": 0.8971962616822431,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8691588785046729,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.875739644970414,
+      "eval_LOCATION_precision": 0.8648648648648649,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8895522388059701,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8862275449101796,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9085365853658537,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04560285061597824,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.908641975308642,
-      "eval_overall_precision": 0.908641975308642,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.5922,
-      "eval_samples_per_second": 287.086,
-      "eval_steps_per_second": 5.066,
+      "eval_loss": 0.04485779628157616,
+      "eval_overall_accuracy": 0.9864678265672466,
+      "eval_overall_f1": 0.9173859432799014,
+      "eval_overall_precision": 0.916256157635468,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 623.573,
+      "eval_steps_per_second": 11.004,
       "step": 7392
     },
     {
       "epoch": 78.0,
-      "grad_norm": 0.6972180008888245,
+      "grad_norm": 0.3859386444091797,
       "learning_rate": 1.1000000000000001e-05,
-      "loss": 0.0397,
+      "loss": 0.0435,
       "step": 7488
     },
     {
       "epoch": 78.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8842729970326408,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8928571428571428,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8975903614457831,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9090909090909091,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045004814863204956,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9111111111111111,
-      "eval_overall_precision": 0.9111111111111111,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5949,
-      "eval_samples_per_second": 285.753,
-      "eval_steps_per_second": 5.043,
+      "eval_loss": 0.044405534863471985,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9211822660098522,
+      "eval_overall_precision": 0.918918918918919,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2722,
+      "eval_samples_per_second": 624.535,
+      "eval_steps_per_second": 11.021,
       "step": 7488
     },
     {
       "epoch": 79.0,
-      "grad_norm": 1.175458312034607,
+      "grad_norm": 1.7985838651657104,
       "learning_rate": 1.05e-05,
-      "loss": 0.0411,
+      "loss": 0.0416,
       "step": 7584
     },
     {
       "epoch": 79.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9023255813953488,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8660714285714286,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8994082840236686,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9101796407185628,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04489603638648987,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.5924,
-      "eval_samples_per_second": 286.953,
-      "eval_steps_per_second": 5.064,
+      "eval_loss": 0.04388272389769554,
+      "eval_overall_accuracy": 0.9861916597624966,
+      "eval_overall_f1": 0.9226993865030675,
+      "eval_overall_precision": 0.9170731707317074,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2714,
+      "eval_samples_per_second": 626.402,
+      "eval_steps_per_second": 11.054,
       "step": 7584
     },
     {
       "epoch": 80.0,
-      "grad_norm": 1.935158610343933,
+      "grad_norm": 0.8757291436195374,
       "learning_rate": 1e-05,
-      "loss": 0.0417,
+      "loss": 0.0414,
       "step": 7680
     },
     {
       "epoch": 80.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8823529411764705,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8961424332344213,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8875739644970414,
-      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_ORGANIZATION_precision": 0.9096385542168675,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04481157660484314,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9102091020910209,
-      "eval_overall_precision": 0.9068627450980392,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.6085,
-      "eval_samples_per_second": 279.37,
-      "eval_steps_per_second": 4.93,
+      "eval_loss": 0.04357600957155228,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9211822660098522,
+      "eval_overall_precision": 0.918918918918919,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2731,
+      "eval_samples_per_second": 622.407,
+      "eval_steps_per_second": 10.984,
       "step": 7680
     },
     {
       "epoch": 81.0,
-      "grad_norm": 0.7792730927467346,
+      "grad_norm": 0.4684378504753113,
       "learning_rate": 9.5e-06,
-      "loss": 0.0423,
+      "loss": 0.043,
       "step": 7776
     },
     {
       "epoch": 81.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8823529411764705,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8875739644970414,
-      "eval_ORGANIZATION_recall": 0.8771929824561403,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04457651078701019,
-      "eval_overall_accuracy": 0.9856393261529964,
-      "eval_overall_f1": 0.9102091020910209,
-      "eval_overall_precision": 0.9068627450980392,
-      "eval_overall_recall": 0.9135802469135802,
-      "eval_runtime": 0.6034,
-      "eval_samples_per_second": 281.758,
-      "eval_steps_per_second": 4.972,
+      "eval_loss": 0.043720223009586334,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2724,
+      "eval_samples_per_second": 624.084,
+      "eval_steps_per_second": 11.013,
       "step": 7776
     },
     {
       "epoch": 82.0,
-      "grad_norm": 1.640648365020752,
+      "grad_norm": 0.9328840970993042,
       "learning_rate": 9e-06,
-      "loss": 0.0434,
+      "loss": 0.0433,
       "step": 7872
     },
     {
       "epoch": 82.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9107981220657277,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8994082840236686,
+      "eval_LOCATION_precision": 0.8818181818181818,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.9020771513353115,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.9101796407185628,
+      "eval_ORGANIZATION_precision": 0.9156626506024096,
       "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04452081397175789,
-      "eval_overall_accuracy": 0.9861916597624966,
-      "eval_overall_f1": 0.9173859432799014,
-      "eval_overall_precision": 0.916256157635468,
-      "eval_overall_recall": 0.9185185185185185,
-      "eval_runtime": 0.6038,
-      "eval_samples_per_second": 281.535,
-      "eval_steps_per_second": 4.968,
+      "eval_loss": 0.04343878850340843,
+      "eval_overall_accuracy": 0.9872963269814968,
+      "eval_overall_f1": 0.9261083743842363,
+      "eval_overall_precision": 0.9238329238329238,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2728,
+      "eval_samples_per_second": 623.154,
+      "eval_steps_per_second": 10.997,
       "step": 7872
     },
     {
       "epoch": 83.0,
-      "grad_norm": 0.8310399651527405,
+      "grad_norm": 0.808896005153656,
       "learning_rate": 8.500000000000002e-06,
-      "loss": 0.0394,
+      "loss": 0.0428,
       "step": 7968
     },
     {
       "epoch": 83.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9023255813953488,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8660714285714286,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.9047619047619047,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9212121212121213,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04489120468497276,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6046,
-      "eval_samples_per_second": 281.181,
-      "eval_steps_per_second": 4.962,
+      "eval_loss": 0.043899569660425186,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9249692496924968,
+      "eval_overall_precision": 0.9215686274509803,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2716,
+      "eval_samples_per_second": 625.96,
+      "eval_steps_per_second": 11.046,
       "step": 7968
     },
     {
       "epoch": 84.0,
-      "grad_norm": 0.9918228387832642,
+      "grad_norm": 0.7827540040016174,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.0413,
+      "loss": 0.0418,
       "step": 8064
     },
     {
       "epoch": 84.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8809523809523809,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.9020771513353115,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.896969696969697,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9156626506024096,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04544991999864578,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.9097651421508034,
-      "eval_overall_precision": 0.9108910891089109,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6004,
-      "eval_samples_per_second": 283.156,
-      "eval_steps_per_second": 4.997,
+      "eval_loss": 0.04345087707042694,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9249692496924968,
+      "eval_overall_precision": 0.9215686274509803,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2707,
+      "eval_samples_per_second": 628.109,
+      "eval_steps_per_second": 11.084,
       "step": 8064
     },
     {
       "epoch": 85.0,
-      "grad_norm": 0.9799830317497253,
+      "grad_norm": 0.7387497425079346,
       "learning_rate": 7.5e-06,
-      "loss": 0.0408,
+      "loss": 0.0416,
       "step": 8160
     },
     {
       "epoch": 85.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8783382789317508,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.891566265060241,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045652709901332855,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.908641975308642,
-      "eval_overall_precision": 0.908641975308642,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.604,
-      "eval_samples_per_second": 281.458,
-      "eval_steps_per_second": 4.967,
+      "eval_loss": 0.04350125417113304,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2731,
+      "eval_samples_per_second": 622.486,
+      "eval_steps_per_second": 10.985,
       "step": 8160
     },
     {
       "epoch": 86.0,
-      "grad_norm": 1.889784574508667,
+      "grad_norm": 0.7260624170303345,
       "learning_rate": 7.000000000000001e-06,
-      "loss": 0.0412,
+      "loss": 0.0413,
       "step": 8256
     },
     {
       "epoch": 86.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.9047619047619047,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9212121212121213,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045468661934137344,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6007,
-      "eval_samples_per_second": 283.007,
-      "eval_steps_per_second": 4.994,
+      "eval_loss": 0.04387590289115906,
+      "eval_overall_accuracy": 0.9872963269814968,
+      "eval_overall_f1": 0.9261083743842363,
+      "eval_overall_precision": 0.9238329238329238,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.2733,
+      "eval_samples_per_second": 622.03,
+      "eval_steps_per_second": 10.977,
       "step": 8256
     },
     {
       "epoch": 87.0,
-      "grad_norm": 1.0306001901626587,
+      "grad_norm": 0.7930686473846436,
       "learning_rate": 6.5000000000000004e-06,
-      "loss": 0.0425,
+      "loss": 0.0423,
       "step": 8352
     },
     {
       "epoch": 87.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8731563421828908,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8895522388059701,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8809523809523809,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9085365853658537,
+      "eval_ORGANIZATION_recall": 0.8713450292397661,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045380450785160065,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.9064039408866995,
-      "eval_overall_precision": 0.9041769041769042,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.607,
-      "eval_samples_per_second": 280.052,
-      "eval_steps_per_second": 4.942,
+      "eval_loss": 0.04400918632745743,
+      "eval_overall_accuracy": 0.9864678265672466,
+      "eval_overall_f1": 0.9185185185185185,
+      "eval_overall_precision": 0.9185185185185185,
+      "eval_overall_recall": 0.9185185185185185,
+      "eval_runtime": 0.2715,
+      "eval_samples_per_second": 626.246,
+      "eval_steps_per_second": 11.051,
       "step": 8352
     },
     {
       "epoch": 88.0,
-      "grad_norm": 1.6095170974731445,
+      "grad_norm": 0.6405452489852905,
       "learning_rate": 6e-06,
-      "loss": 0.0404,
+      "loss": 0.0409,
       "step": 8448
     },
     {
       "epoch": 88.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9107981220657277,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8818181818181818,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045171529054641724,
-      "eval_overall_accuracy": 0.9850869925434963,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5928,
-      "eval_samples_per_second": 286.772,
-      "eval_steps_per_second": 5.061,
+      "eval_loss": 0.04390283301472664,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9234567901234567,
+      "eval_overall_precision": 0.9234567901234568,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2725,
+      "eval_samples_per_second": 623.922,
+      "eval_steps_per_second": 11.01,
       "step": 8448
     },
     {
       "epoch": 89.0,
-      "grad_norm": 1.9292471408843994,
+      "grad_norm": 0.8073258996009827,
       "learning_rate": 5.500000000000001e-06,
-      "loss": 0.0403,
+      "loss": 0.0419,
       "step": 8544
     },
     {
       "epoch": 89.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9023255813953488,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.875739644970414,
+      "eval_LOCATION_precision": 0.8660714285714286,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.9047619047619047,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8862275449101796,
-      "eval_ORGANIZATION_recall": 0.8654970760233918,
+      "eval_ORGANIZATION_precision": 0.9212121212121213,
+      "eval_ORGANIZATION_recall": 0.8888888888888888,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04589909315109253,
-      "eval_overall_accuracy": 0.9842584921292461,
-      "eval_overall_f1": 0.9075215782983971,
-      "eval_overall_precision": 0.9064039408866995,
-      "eval_overall_recall": 0.908641975308642,
-      "eval_runtime": 0.6049,
-      "eval_samples_per_second": 281.046,
-      "eval_steps_per_second": 4.96,
+      "eval_loss": 0.043653201311826706,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9249692496924968,
+      "eval_overall_precision": 0.9215686274509803,
+      "eval_overall_recall": 0.928395061728395,
+      "eval_runtime": 0.271,
+      "eval_samples_per_second": 627.261,
+      "eval_steps_per_second": 11.069,
       "step": 8544
     },
     {
       "epoch": 90.0,
-      "grad_norm": 0.9496196508407593,
+      "grad_norm": 0.8635096549987793,
       "learning_rate": 5e-06,
-      "loss": 0.0395,
+      "loss": 0.0424,
       "step": 8640
     },
     {
       "epoch": 90.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9023255813953488,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8660714285714286,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04514380544424057,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6057,
-      "eval_samples_per_second": 280.676,
-      "eval_steps_per_second": 4.953,
+      "eval_loss": 0.043814968317747116,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.922509225092251,
+      "eval_overall_precision": 0.9191176470588235,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2729,
+      "eval_samples_per_second": 622.855,
+      "eval_steps_per_second": 10.992,
       "step": 8640
     },
     {
       "epoch": 91.0,
-      "grad_norm": 1.498864769935608,
+      "grad_norm": 0.7420637607574463,
       "learning_rate": 4.5e-06,
-      "loss": 0.0413,
+      "loss": 0.0419,
       "step": 8736
     },
     {
       "epoch": 91.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04504216089844704,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6,
-      "eval_samples_per_second": 283.336,
-      "eval_steps_per_second": 5.0,
+      "eval_loss": 0.04388648644089699,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2731,
+      "eval_samples_per_second": 622.382,
+      "eval_steps_per_second": 10.983,
       "step": 8736
     },
     {
       "epoch": 92.0,
-      "grad_norm": 0.8939065337181091,
+      "grad_norm": 0.2291034311056137,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.039,
+      "loss": 0.0427,
       "step": 8832
     },
     {
       "epoch": 92.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04519949480891228,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6015,
-      "eval_samples_per_second": 282.646,
-      "eval_steps_per_second": 4.988,
+      "eval_loss": 0.044274233281612396,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2719,
+      "eval_samples_per_second": 625.194,
+      "eval_steps_per_second": 11.033,
       "step": 8832
     },
     {
       "epoch": 93.0,
-      "grad_norm": 76.55709838867188,
+      "grad_norm": 0.9127195477485657,
       "learning_rate": 3.5000000000000004e-06,
-      "loss": 0.0402,
+      "loss": 0.0397,
       "step": 8928
     },
     {
       "epoch": 93.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045382946729660034,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5985,
-      "eval_samples_per_second": 284.026,
-      "eval_steps_per_second": 5.012,
+      "eval_loss": 0.04379863291978836,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9223181257706535,
+      "eval_overall_precision": 0.9211822660098522,
+      "eval_overall_recall": 0.9234567901234568,
+      "eval_runtime": 0.2732,
+      "eval_samples_per_second": 622.226,
+      "eval_steps_per_second": 10.98,
       "step": 8928
     },
     {
       "epoch": 94.0,
-      "grad_norm": 0.24397915601730347,
+      "grad_norm": 1.1270931959152222,
       "learning_rate": 3e-06,
-      "loss": 0.0397,
+      "loss": 0.0414,
       "step": 9024
     },
     {
       "epoch": 94.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045313529670238495,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5887,
-      "eval_samples_per_second": 288.767,
-      "eval_steps_per_second": 5.096,
+      "eval_loss": 0.043688420206308365,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9236453201970444,
+      "eval_overall_precision": 0.9213759213759214,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2716,
+      "eval_samples_per_second": 625.967,
+      "eval_steps_per_second": 11.046,
       "step": 9024
     },
     {
       "epoch": 95.0,
-      "grad_norm": 0.29766303300857544,
+      "grad_norm": 0.35525602102279663,
       "learning_rate": 2.5e-06,
-      "loss": 0.0409,
+      "loss": 0.0401,
       "step": 9120
     },
     {
       "epoch": 95.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9107981220657277,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8818181818181818,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045457012951374054,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5948,
-      "eval_samples_per_second": 285.825,
-      "eval_steps_per_second": 5.044,
+      "eval_loss": 0.043763358145952225,
+      "eval_overall_accuracy": 0.9872963269814968,
+      "eval_overall_f1": 0.9247842170160296,
+      "eval_overall_precision": 0.9236453201970444,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2719,
+      "eval_samples_per_second": 625.255,
+      "eval_steps_per_second": 11.034,
       "step": 9120
     },
     {
       "epoch": 96.0,
-      "grad_norm": 1.5634249448776245,
+      "grad_norm": 0.21041779220104218,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.039,
+      "loss": 0.0415,
       "step": 9216
     },
     {
       "epoch": 96.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9014084507042254,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8727272727272727,
+      "eval_LOCATION_recall": 0.9320388349514563,
+      "eval_ORGANIZATION_f1": 0.8955223880597015,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9146341463414634,
+      "eval_ORGANIZATION_recall": 0.8771929824561403,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.0454791858792305,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5934,
-      "eval_samples_per_second": 286.492,
-      "eval_steps_per_second": 5.056,
+      "eval_loss": 0.043890830129384995,
+      "eval_overall_accuracy": 0.9867439933719967,
+      "eval_overall_f1": 0.9209876543209876,
+      "eval_overall_precision": 0.9209876543209876,
+      "eval_overall_recall": 0.9209876543209876,
+      "eval_runtime": 0.2726,
+      "eval_samples_per_second": 623.718,
+      "eval_steps_per_second": 11.007,
       "step": 9216
     },
     {
       "epoch": 97.0,
-      "grad_norm": 0.9605298042297363,
+      "grad_norm": 0.8473381996154785,
       "learning_rate": 1.5e-06,
-      "loss": 0.0402,
+      "loss": 0.0404,
       "step": 9312
     },
     {
       "epoch": 97.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9107981220657277,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8739002932551321,
+      "eval_LOCATION_precision": 0.8818181818181818,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8764705882352941,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.0456123985350132,
-      "eval_overall_accuracy": 0.9845346589339962,
-      "eval_overall_f1": 0.9066339066339066,
-      "eval_overall_precision": 0.902200488997555,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6745,
-      "eval_samples_per_second": 252.035,
-      "eval_steps_per_second": 4.448,
+      "eval_loss": 0.043740756809711456,
+      "eval_overall_accuracy": 0.9872963269814968,
+      "eval_overall_f1": 0.9247842170160296,
+      "eval_overall_precision": 0.9236453201970444,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2732,
+      "eval_samples_per_second": 622.321,
+      "eval_steps_per_second": 10.982,
       "step": 9312
     },
     {
       "epoch": 98.0,
-      "grad_norm": 1.7611069679260254,
+      "grad_norm": 1.3456790447235107,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 0.0382,
+      "loss": 0.0418,
       "step": 9408
     },
     {
       "epoch": 98.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04547755792737007,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5989,
-      "eval_samples_per_second": 283.847,
-      "eval_steps_per_second": 5.009,
+      "eval_loss": 0.04376491159200668,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9236453201970444,
+      "eval_overall_precision": 0.9213759213759214,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2717,
+      "eval_samples_per_second": 625.746,
+      "eval_steps_per_second": 11.043,
       "step": 9408
     },
     {
       "epoch": 99.0,
-      "grad_norm": 0.4898216426372528,
+      "grad_norm": 2.012141704559326,
       "learning_rate": 5.000000000000001e-07,
-      "loss": 0.0392,
+      "loss": 0.0388,
       "step": 9504
     },
     {
       "epoch": 99.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.04550032317638397,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.5984,
-      "eval_samples_per_second": 284.085,
-      "eval_steps_per_second": 5.013,
+      "eval_loss": 0.043746430426836014,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9236453201970444,
+      "eval_overall_precision": 0.9213759213759214,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2735,
+      "eval_samples_per_second": 621.47,
+      "eval_steps_per_second": 10.967,
       "step": 9504
     },
     {
       "epoch": 100.0,
-      "grad_norm": 1.0116997957229614,
+      "grad_norm": 0.3253363370895386,
       "learning_rate": 0.0,
-      "loss": 0.0412,
+      "loss": 0.0397,
       "step": 9600
     },
     {
       "epoch": 100.0,
-      "eval_LOCATION_f1": 0.8815165876777251,
+      "eval_LOCATION_f1": 0.9065420560747663,
       "eval_LOCATION_number": 103,
-      "eval_LOCATION_precision": 0.8611111111111112,
-      "eval_LOCATION_recall": 0.9029126213592233,
-      "eval_ORGANIZATION_f1": 0.8764705882352941,
+      "eval_LOCATION_precision": 0.8738738738738738,
+      "eval_LOCATION_recall": 0.941747572815534,
+      "eval_ORGANIZATION_f1": 0.8988095238095237,
       "eval_ORGANIZATION_number": 171,
-      "eval_ORGANIZATION_precision": 0.8816568047337278,
-      "eval_ORGANIZATION_recall": 0.8713450292397661,
+      "eval_ORGANIZATION_precision": 0.9151515151515152,
+      "eval_ORGANIZATION_recall": 0.8830409356725146,
       "eval_PERSON_f1": 0.9694656488549618,
       "eval_PERSON_number": 131,
       "eval_PERSON_precision": 0.9694656488549618,
       "eval_PERSON_recall": 0.9694656488549618,
-      "eval_loss": 0.045390497893095016,
-      "eval_overall_accuracy": 0.9848108257387462,
-      "eval_overall_f1": 0.9077490774907748,
-      "eval_overall_precision": 0.9044117647058824,
-      "eval_overall_recall": 0.9111111111111111,
-      "eval_runtime": 0.6032,
-      "eval_samples_per_second": 281.85,
-      "eval_steps_per_second": 4.974,
+      "eval_loss": 0.04369127005338669,
+      "eval_overall_accuracy": 0.9870201601767468,
+      "eval_overall_f1": 0.9236453201970444,
+      "eval_overall_precision": 0.9213759213759214,
+      "eval_overall_recall": 0.9259259259259259,
+      "eval_runtime": 0.2733,
+      "eval_samples_per_second": 622.123,
+      "eval_steps_per_second": 10.979,
       "step": 9600
     },
     {
       "epoch": 100.0,
       "step": 9600,
       "total_flos": 3889996718638620.0,
-      "train_loss": 0.09181262865662575,
-      "train_runtime": 1037.4922,
-      "train_samples_per_second": 147.567,
-      "train_steps_per_second": 9.253
+      "train_loss": 0.09361960942546527,
+      "train_runtime": 547.4358,
+      "train_samples_per_second": 279.667,
+      "train_steps_per_second": 17.536
     }
   ],
   "logging_steps": 500,