diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,5488 @@
+{
+  "best_metric": 0.9738785407252287,
+  "best_model_checkpoint": "drive/MyDrive/albertina/checkpoint-9000",
+  "epoch": 0.8024966562639322,
+  "global_step": 9000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 0.7187,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 0.7199,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 0.6146,
+      "step": 30
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 0.6195,
+      "step": 40
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 0.6245,
+      "step": 50
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 0.6241,
+      "step": 60
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 0.5657,
+      "step": 70
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 0.4948,
+      "step": 80
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.8000000000000001e-06,
+      "loss": 0.5167,
+      "step": 90
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.4677,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.2e-06,
+      "loss": 0.4529,
+      "step": 110
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 0.4148,
+      "step": 120
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.6e-06,
+      "loss": 0.3799,
+      "step": 130
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 0.3091,
+      "step": 140
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3e-06,
+      "loss": 0.3131,
+      "step": 150
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 0.2855,
+      "step": 160
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 0.2601,
+      "step": 170
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 0.193,
+      "step": 180
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.8000000000000005e-06,
+      "loss": 0.2041,
+      "step": 190
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.326,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.2000000000000004e-06,
+      "loss": 0.1665,
+      "step": 210
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.4e-06,
+      "loss": 0.2885,
+      "step": 220
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 0.2399,
+      "step": 230
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 0.2109,
+      "step": 240
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5e-06,
+      "loss": 0.2259,
+      "step": 250
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.2e-06,
+      "loss": 0.2323,
+      "step": 260
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.400000000000001e-06,
+      "loss": 0.1955,
+      "step": 270
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 0.1765,
+      "step": 280
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.8e-06,
+      "loss": 0.1722,
+      "step": 290
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-06,
+      "loss": 0.1648,
+      "step": 300
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.200000000000001e-06,
+      "loss": 0.1617,
+      "step": 310
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 0.1687,
+      "step": 320
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.600000000000001e-06,
+      "loss": 0.1727,
+      "step": 330
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6.800000000000001e-06,
+      "loss": 0.2064,
+      "step": 340
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7e-06,
+      "loss": 0.1663,
+      "step": 350
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 0.1032,
+      "step": 360
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7.4e-06,
+      "loss": 0.1592,
+      "step": 370
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7.600000000000001e-06,
+      "loss": 0.2606,
+      "step": 380
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7.800000000000002e-06,
+      "loss": 0.0709,
+      "step": 390
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.117,
+      "step": 400
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.2e-06,
+      "loss": 0.1836,
+      "step": 410
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 0.0998,
+      "step": 420
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.6e-06,
+      "loss": 0.1288,
+      "step": 430
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 8.8e-06,
+      "loss": 0.1877,
+      "step": 440
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9e-06,
+      "loss": 0.1695,
+      "step": 450
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.200000000000002e-06,
+      "loss": 0.1736,
+      "step": 460
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.4e-06,
+      "loss": 0.195,
+      "step": 470
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 0.1599,
+      "step": 480
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 9.800000000000001e-06,
+      "loss": 0.0596,
+      "step": 490
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1e-05,
+      "loss": 0.0599,
+      "step": 500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.998200629779578e-06,
+      "loss": 0.2157,
+      "step": 510
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.996401259559155e-06,
+      "loss": 0.1006,
+      "step": 520
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.994601889338731e-06,
+      "loss": 0.2082,
+      "step": 530
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.99280251911831e-06,
+      "loss": 0.2305,
+      "step": 540
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.991003148897887e-06,
+      "loss": 0.1995,
+      "step": 550
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.989203778677464e-06,
+      "loss": 0.1551,
+      "step": 560
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.98740440845704e-06,
+      "loss": 0.0971,
+      "step": 570
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.985605038236617e-06,
+      "loss": 0.2007,
+      "step": 580
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.983805668016196e-06,
+      "loss": 0.1312,
+      "step": 590
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.982006297795773e-06,
+      "loss": 0.1931,
+      "step": 600
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 9.98020692757535e-06,
+      "loss": 0.223,
+      "step": 610
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.978407557354927e-06,
+      "loss": 0.3252,
+      "step": 620
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.976608187134503e-06,
+      "loss": 0.0981,
+      "step": 630
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.97480881691408e-06,
+      "loss": 0.1815,
+      "step": 640
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.973009446693657e-06,
+      "loss": 0.1782,
+      "step": 650
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.971210076473236e-06,
+      "loss": 0.1804,
+      "step": 660
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.969410706252813e-06,
+      "loss": 0.1271,
+      "step": 670
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.96761133603239e-06,
+      "loss": 0.1316,
+      "step": 680
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.965811965811966e-06,
+      "loss": 0.2092,
+      "step": 690
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.964012595591543e-06,
+      "loss": 0.0782,
+      "step": 700
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.962213225371122e-06,
+      "loss": 0.2433,
+      "step": 710
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.960413855150699e-06,
+      "loss": 0.1883,
+      "step": 720
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.958614484930275e-06,
+      "loss": 0.032,
+      "step": 730
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.956815114709852e-06,
+      "loss": 0.2181,
+      "step": 740
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.955015744489429e-06,
+      "loss": 0.1828,
+      "step": 750
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.953216374269008e-06,
+      "loss": 0.1206,
+      "step": 760
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.951417004048583e-06,
+      "loss": 0.1253,
+      "step": 770
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.949617633828161e-06,
+      "loss": 0.1047,
+      "step": 780
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.947818263607738e-06,
+      "loss": 0.1174,
+      "step": 790
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.946018893387315e-06,
+      "loss": 0.3025,
+      "step": 800
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.944219523166892e-06,
+      "loss": 0.0581,
+      "step": 810
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.942420152946469e-06,
+      "loss": 0.2664,
+      "step": 820
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.940620782726047e-06,
+      "loss": 0.1218,
+      "step": 830
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.938821412505624e-06,
+      "loss": 0.1708,
+      "step": 840
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.937022042285201e-06,
+      "loss": 0.1626,
+      "step": 850
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.935222672064778e-06,
+      "loss": 0.0553,
+      "step": 860
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.933423301844355e-06,
+      "loss": 0.1209,
+      "step": 870
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.931623931623933e-06,
+      "loss": 0.11,
+      "step": 880
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.929824561403509e-06,
+      "loss": 0.0945,
+      "step": 890
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.928025191183087e-06,
+      "loss": 0.2105,
+      "step": 900
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.926225820962664e-06,
+      "loss": 0.1548,
+      "step": 910
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.92442645074224e-06,
+      "loss": 0.1819,
+      "step": 920
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.922627080521818e-06,
+      "loss": 0.1461,
+      "step": 930
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.920827710301395e-06,
+      "loss": 0.1917,
+      "step": 940
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.919028340080973e-06,
+      "loss": 0.0796,
+      "step": 950
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.91722896986055e-06,
+      "loss": 0.1768,
+      "step": 960
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.915429599640127e-06,
+      "loss": 0.1726,
+      "step": 970
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.913630229419704e-06,
+      "loss": 0.1244,
+      "step": 980
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.91183085919928e-06,
+      "loss": 0.0765,
+      "step": 990
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.910031488978859e-06,
+      "loss": 0.1842,
+      "step": 1000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.908232118758436e-06,
+      "loss": 0.1192,
+      "step": 1010
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.906432748538013e-06,
+      "loss": 0.132,
+      "step": 1020
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.90463337831759e-06,
+      "loss": 0.1347,
+      "step": 1030
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.902834008097167e-06,
+      "loss": 0.0591,
+      "step": 1040
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.901034637876743e-06,
+      "loss": 0.1588,
+      "step": 1050
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 9.89923526765632e-06,
+      "loss": 0.1273,
+      "step": 1060
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.897435897435899e-06,
+      "loss": 0.0527,
+      "step": 1070
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.895636527215476e-06,
+      "loss": 0.1438,
+      "step": 1080
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.893837156995053e-06,
+      "loss": 0.1085,
+      "step": 1090
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.89203778677463e-06,
+      "loss": 0.0793,
+      "step": 1100
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.890238416554206e-06,
+      "loss": 0.1617,
+      "step": 1110
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.888439046333785e-06,
+      "loss": 0.0164,
+      "step": 1120
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.886639676113362e-06,
+      "loss": 0.2053,
+      "step": 1130
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.884840305892939e-06,
+      "loss": 0.1774,
+      "step": 1140
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.883040935672515e-06,
+      "loss": 0.1195,
+      "step": 1150
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.881241565452092e-06,
+      "loss": 0.0487,
+      "step": 1160
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.879442195231669e-06,
+      "loss": 0.1629,
+      "step": 1170
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.877642825011246e-06,
+      "loss": 0.0964,
+      "step": 1180
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.875843454790825e-06,
+      "loss": 0.1443,
+      "step": 1190
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.874044084570401e-06,
+      "loss": 0.1089,
+      "step": 1200
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.872244714349978e-06,
+      "loss": 0.0233,
+      "step": 1210
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.870445344129555e-06,
+      "loss": 0.0712,
+      "step": 1220
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.868645973909132e-06,
+      "loss": 0.2537,
+      "step": 1230
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.86684660368871e-06,
+      "loss": 0.2621,
+      "step": 1240
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.865047233468287e-06,
+      "loss": 0.1302,
+      "step": 1250
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.863247863247864e-06,
+      "loss": 0.136,
+      "step": 1260
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.861448493027441e-06,
+      "loss": 0.146,
+      "step": 1270
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.859649122807018e-06,
+      "loss": 0.0938,
+      "step": 1280
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.857849752586597e-06,
+      "loss": 0.0718,
+      "step": 1290
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.856050382366172e-06,
+      "loss": 0.1728,
+      "step": 1300
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.85425101214575e-06,
+      "loss": 0.2543,
+      "step": 1310
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.852451641925327e-06,
+      "loss": 0.1327,
+      "step": 1320
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.850652271704904e-06,
+      "loss": 0.1058,
+      "step": 1330
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.84885290148448e-06,
+      "loss": 0.13,
+      "step": 1340
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.847053531264058e-06,
+      "loss": 0.1322,
+      "step": 1350
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.845254161043636e-06,
+      "loss": 0.1909,
+      "step": 1360
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.843454790823213e-06,
+      "loss": 0.078,
+      "step": 1370
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.84165542060279e-06,
+      "loss": 0.1836,
+      "step": 1380
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.839856050382367e-06,
+      "loss": 0.1491,
+      "step": 1390
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.838056680161944e-06,
+      "loss": 0.139,
+      "step": 1400
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.836257309941522e-06,
+      "loss": 0.0737,
+      "step": 1410
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.834457939721097e-06,
+      "loss": 0.1696,
+      "step": 1420
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.832658569500676e-06,
+      "loss": 0.2328,
+      "step": 1430
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.830859199280253e-06,
+      "loss": 0.1342,
+      "step": 1440
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.82905982905983e-06,
+      "loss": 0.1243,
+      "step": 1450
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.827260458839407e-06,
+      "loss": 0.1969,
+      "step": 1460
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.825461088618983e-06,
+      "loss": 0.2298,
+      "step": 1470
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.823661718398562e-06,
+      "loss": 0.1167,
+      "step": 1480
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.821862348178139e-06,
+      "loss": 0.123,
+      "step": 1490
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.820062977957716e-06,
+      "loss": 0.1128,
+      "step": 1500
+    },
+    {
+      "epoch": 0.13,
+      "eval_accuracy": 0.9694931312174325,
+      "eval_f1": 0.94606967396159,
+      "eval_loss": 0.09655023366212845,
+      "eval_precision": 0.9585926009729607,
+      "eval_recall": 0.9338697233550094,
+      "eval_runtime": 436.9466,
+      "eval_samples_per_second": 72.469,
+      "eval_steps_per_second": 4.531,
+      "step": 1500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.818263607737293e-06,
+      "loss": 0.0397,
+      "step": 1510
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.81646423751687e-06,
+      "loss": 0.1572,
+      "step": 1520
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.814664867296448e-06,
+      "loss": 0.1214,
+      "step": 1530
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.812865497076025e-06,
+      "loss": 0.135,
+      "step": 1540
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.811066126855602e-06,
+      "loss": 0.3403,
+      "step": 1550
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.809266756635179e-06,
+      "loss": 0.0645,
+      "step": 1560
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.807467386414755e-06,
+      "loss": 0.1162,
+      "step": 1570
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.805668016194332e-06,
+      "loss": 0.065,
+      "step": 1580
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.803868645973909e-06,
+      "loss": 0.0923,
+      "step": 1590
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.802069275753488e-06,
+      "loss": 0.2101,
+      "step": 1600
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.800269905533065e-06,
+      "loss": 0.1123,
+      "step": 1610
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.798470535312641e-06,
+      "loss": 0.2323,
+      "step": 1620
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.796671165092218e-06,
+      "loss": 0.0653,
+      "step": 1630
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.794871794871795e-06,
+      "loss": 0.1639,
+      "step": 1640
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.793072424651374e-06,
+      "loss": 0.0505,
+      "step": 1650
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.79127305443095e-06,
+      "loss": 0.1409,
+      "step": 1660
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.789473684210527e-06,
+      "loss": 0.1419,
+      "step": 1670
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.787674313990104e-06,
+      "loss": 0.1625,
+      "step": 1680
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.785874943769681e-06,
+      "loss": 0.0771,
+      "step": 1690
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.784075573549258e-06,
+      "loss": 0.0929,
+      "step": 1700
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.782276203328835e-06,
+      "loss": 0.095,
+      "step": 1710
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.780476833108413e-06,
+      "loss": 0.1673,
+      "step": 1720
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.77867746288799e-06,
+      "loss": 0.1376,
+      "step": 1730
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.776878092667567e-06,
+      "loss": 0.0783,
+      "step": 1740
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.775078722447144e-06,
+      "loss": 0.1428,
+      "step": 1750
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.77327935222672e-06,
+      "loss": 0.0576,
+      "step": 1760
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.7714799820063e-06,
+      "loss": 0.0739,
+      "step": 1770
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.769680611785876e-06,
+      "loss": 0.09,
+      "step": 1780
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.767881241565453e-06,
+      "loss": 0.0987,
+      "step": 1790
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.76608187134503e-06,
+      "loss": 0.1613,
+      "step": 1800
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.764282501124607e-06,
+      "loss": 0.058,
+      "step": 1810
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.762483130904185e-06,
+      "loss": 0.218,
+      "step": 1820
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.76068376068376e-06,
+      "loss": 0.1083,
+      "step": 1830
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.758884390463339e-06,
+      "loss": 0.1274,
+      "step": 1840
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.757085020242916e-06,
+      "loss": 0.1016,
+      "step": 1850
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.755285650022493e-06,
+      "loss": 0.1442,
+      "step": 1860
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.75348627980207e-06,
+      "loss": 0.0771,
+      "step": 1870
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.751686909581647e-06,
+      "loss": 0.0832,
+      "step": 1880
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.749887539361225e-06,
+      "loss": 0.1264,
+      "step": 1890
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.748088169140802e-06,
+      "loss": 0.1182,
+      "step": 1900
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.746288798920379e-06,
+      "loss": 0.1582,
+      "step": 1910
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.744489428699956e-06,
+      "loss": 0.1256,
+      "step": 1920
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.742690058479533e-06,
+      "loss": 0.1081,
+      "step": 1930
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.740890688259111e-06,
+      "loss": 0.0922,
+      "step": 1940
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.739091318038686e-06,
+      "loss": 0.1155,
+      "step": 1950
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.737291947818265e-06,
+      "loss": 0.0559,
+      "step": 1960
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.735492577597842e-06,
+      "loss": 0.1008,
+      "step": 1970
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.733693207377419e-06,
+      "loss": 0.1123,
+      "step": 1980
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.731893837156995e-06,
+      "loss": 0.1095,
+      "step": 1990
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.730094466936572e-06,
+      "loss": 0.1269,
+      "step": 2000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.72829509671615e-06,
+      "loss": 0.1702,
+      "step": 2010
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.726495726495728e-06,
+      "loss": 0.1762,
+      "step": 2020
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.724696356275305e-06,
+      "loss": 0.0528,
+      "step": 2030
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.722896986054881e-06,
+      "loss": 0.0912,
+      "step": 2040
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.721097615834458e-06,
+      "loss": 0.0528,
+      "step": 2050
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.719298245614037e-06,
+      "loss": 0.1277,
+      "step": 2060
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.717498875393614e-06,
+      "loss": 0.1685,
+      "step": 2070
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.71569950517319e-06,
+      "loss": 0.0825,
+      "step": 2080
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.713900134952767e-06,
+      "loss": 0.1238,
+      "step": 2090
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.712100764732344e-06,
+      "loss": 0.0815,
+      "step": 2100
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.710301394511921e-06,
+      "loss": 0.0942,
+      "step": 2110
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.708502024291498e-06,
+      "loss": 0.0789,
+      "step": 2120
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.706702654071076e-06,
+      "loss": 0.1439,
+      "step": 2130
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.704903283850653e-06,
+      "loss": 0.1,
+      "step": 2140
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.70310391363023e-06,
+      "loss": 0.0548,
+      "step": 2150
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.701304543409807e-06,
+      "loss": 0.1491,
+      "step": 2160
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.699505173189384e-06,
+      "loss": 0.063,
+      "step": 2170
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.697705802968962e-06,
+      "loss": 0.2628,
+      "step": 2180
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.69590643274854e-06,
+      "loss": 0.2376,
+      "step": 2190
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.694107062528116e-06,
+      "loss": 0.1094,
+      "step": 2200
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.692307692307693e-06,
+      "loss": 0.143,
+      "step": 2210
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.69050832208727e-06,
+      "loss": 0.1503,
+      "step": 2220
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.688708951866847e-06,
+      "loss": 0.1998,
+      "step": 2230
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.686909581646424e-06,
+      "loss": 0.0649,
+      "step": 2240
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.685110211426002e-06,
+      "loss": 0.024,
+      "step": 2250
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.683310841205579e-06,
+      "loss": 0.119,
+      "step": 2260
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.681511470985156e-06,
+      "loss": 0.228,
+      "step": 2270
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.679712100764733e-06,
+      "loss": 0.1202,
+      "step": 2280
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.67791273054431e-06,
+      "loss": 0.053,
+      "step": 2290
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.676113360323888e-06,
+      "loss": 0.1156,
+      "step": 2300
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.674313990103465e-06,
+      "loss": 0.1197,
+      "step": 2310
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.672514619883042e-06,
+      "loss": 0.0872,
+      "step": 2320
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.670715249662619e-06,
+      "loss": 0.1937,
+      "step": 2330
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.668915879442196e-06,
+      "loss": 0.172,
+      "step": 2340
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.667116509221774e-06,
+      "loss": 0.1065,
+      "step": 2350
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.66531713900135e-06,
+      "loss": 0.12,
+      "step": 2360
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.663517768780928e-06,
+      "loss": 0.094,
+      "step": 2370
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.661718398560505e-06,
+      "loss": 0.0623,
+      "step": 2380
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.659919028340082e-06,
+      "loss": 0.1098,
+      "step": 2390
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.658119658119659e-06,
+      "loss": 0.1679,
+      "step": 2400
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.656320287899235e-06,
+      "loss": 0.1222,
+      "step": 2410
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.654520917678814e-06,
+      "loss": 0.0483,
+      "step": 2420
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.65272154745839e-06,
+      "loss": 0.1289,
+      "step": 2430
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.650922177237968e-06,
+      "loss": 0.0802,
+      "step": 2440
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.649122807017545e-06,
+      "loss": 0.0691,
+      "step": 2450
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.647323436797121e-06,
+      "loss": 0.2148,
+      "step": 2460
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.6455240665767e-06,
+      "loss": 0.1005,
+      "step": 2470
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.643724696356275e-06,
+      "loss": 0.1019,
+      "step": 2480
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.641925326135854e-06,
+      "loss": 0.1362,
+      "step": 2490
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.64012595591543e-06,
+      "loss": 0.0445,
+      "step": 2500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.638326585695007e-06,
+      "loss": 0.1021,
+      "step": 2510
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.636527215474584e-06,
+      "loss": 0.2398,
+      "step": 2520
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.634727845254161e-06,
+      "loss": 0.1228,
+      "step": 2530
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.63292847503374e-06,
+      "loss": 0.1529,
+      "step": 2540
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.631129104813316e-06,
+      "loss": 0.0813,
+      "step": 2550
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.629329734592893e-06,
+      "loss": 0.0542,
+      "step": 2560
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.62753036437247e-06,
+      "loss": 0.1951,
+      "step": 2570
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.625730994152047e-06,
+      "loss": 0.0304,
+      "step": 2580
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.623931623931626e-06,
+      "loss": 0.1102,
+      "step": 2590
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.6221322537112e-06,
+      "loss": 0.0727,
+      "step": 2600
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.62033288349078e-06,
+      "loss": 0.136,
+      "step": 2610
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.618533513270356e-06,
+      "loss": 0.1733,
+      "step": 2620
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.616734143049933e-06,
+      "loss": 0.043,
+      "step": 2630
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.61493477282951e-06,
+      "loss": 0.1242,
+      "step": 2640
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.613135402609087e-06,
+      "loss": 0.0807,
+      "step": 2650
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.611336032388665e-06,
+      "loss": 0.084,
+      "step": 2660
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.609536662168242e-06,
+      "loss": 0.0175,
+      "step": 2670
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.607737291947819e-06,
+      "loss": 0.0686,
+      "step": 2680
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.605937921727396e-06,
+      "loss": 0.0984,
+      "step": 2690
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.604138551506973e-06,
+      "loss": 0.0717,
+      "step": 2700
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.602339181286551e-06,
+      "loss": 0.0323,
+      "step": 2710
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.600539811066128e-06,
+      "loss": 0.1174,
+      "step": 2720
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.598740440845705e-06,
+      "loss": 0.2252,
+      "step": 2730
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.596941070625282e-06,
+      "loss": 0.1463,
+      "step": 2740
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.595141700404859e-06,
+      "loss": 0.1296,
+      "step": 2750
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.593342330184436e-06,
+      "loss": 0.0498,
+      "step": 2760
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.591542959964013e-06,
+      "loss": 0.0488,
+      "step": 2770
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.589743589743591e-06,
+      "loss": 0.1701,
+      "step": 2780
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.587944219523168e-06,
+      "loss": 0.0986,
+      "step": 2790
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.586144849302745e-06,
+      "loss": 0.0605,
+      "step": 2800
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.584345479082322e-06,
+      "loss": 0.0779,
+      "step": 2810
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.582546108861898e-06,
+      "loss": 0.1724,
+      "step": 2820
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.580746738641477e-06,
+      "loss": 0.0229,
+      "step": 2830
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.578947368421054e-06,
+      "loss": 0.1895,
+      "step": 2840
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.57714799820063e-06,
+      "loss": 0.1627,
+      "step": 2850
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.575348627980208e-06,
+      "loss": 0.061,
+      "step": 2860
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.573549257759784e-06,
+      "loss": 0.0394,
+      "step": 2870
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.571749887539361e-06,
+      "loss": 0.0101,
+      "step": 2880
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.569950517318938e-06,
+      "loss": 0.1223,
+      "step": 2890
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.568151147098517e-06,
+      "loss": 0.0859,
+      "step": 2900
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.566351776878094e-06,
+      "loss": 0.0245,
+      "step": 2910
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.56455240665767e-06,
+      "loss": 0.077,
+      "step": 2920
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.562753036437247e-06,
+      "loss": 0.1988,
+      "step": 2930
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.560953666216824e-06,
+      "loss": 0.1803,
+      "step": 2940
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.559154295996403e-06,
+      "loss": 0.0734,
+      "step": 2950
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.55735492577598e-06,
+      "loss": 0.0309,
+      "step": 2960
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.555555555555556e-06,
+      "loss": 0.0958,
+      "step": 2970
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.553756185335133e-06,
+      "loss": 0.1951,
+      "step": 2980
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.55195681511471e-06,
+      "loss": 0.0235,
+      "step": 2990
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.550157444894289e-06,
+      "loss": 0.0854,
+      "step": 3000
+    },
+    {
+      "epoch": 0.27,
+      "eval_accuracy": 0.9742618032528028,
+      "eval_f1": 0.9549798375959787,
+      "eval_loss": 0.1135290339589119,
+      "eval_precision": 0.9572535991140643,
+      "eval_recall": 0.9527168521988317,
+      "eval_runtime": 437.2717,
+      "eval_samples_per_second": 72.415,
+      "eval_steps_per_second": 4.528,
+      "step": 3000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.548358074673864e-06,
+      "loss": 0.0418,
+      "step": 3010
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.546558704453442e-06,
+      "loss": 0.0694,
+      "step": 3020
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.54475933423302e-06,
+      "loss": 0.048,
+      "step": 3030
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.542959964012596e-06,
+      "loss": 0.1657,
+      "step": 3040
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.541160593792173e-06,
+      "loss": 0.0778,
+      "step": 3050
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.53936122357175e-06,
+      "loss": 0.1125,
+      "step": 3060
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.537561853351328e-06,
+      "loss": 0.0662,
+      "step": 3070
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.535762483130905e-06,
+      "loss": 0.1191,
+      "step": 3080
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.533963112910482e-06,
+      "loss": 0.171,
+      "step": 3090
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.532163742690059e-06,
+      "loss": 0.0703,
+      "step": 3100
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.530364372469636e-06,
+      "loss": 0.138,
+      "step": 3110
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.528565002249214e-06,
+      "loss": 0.1945,
+      "step": 3120
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.52676563202879e-06,
+      "loss": 0.1241,
+      "step": 3130
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.524966261808368e-06,
+      "loss": 0.1185,
+      "step": 3140
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.523166891587945e-06,
+      "loss": 0.0739,
+      "step": 3150
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.521367521367522e-06,
+      "loss": 0.0361,
+      "step": 3160
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.519568151147099e-06,
+      "loss": 0.1103,
+      "step": 3170
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.517768780926676e-06,
+      "loss": 0.1141,
+      "step": 3180
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.515969410706254e-06,
+      "loss": 0.0741,
+      "step": 3190
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.514170040485831e-06,
+      "loss": 0.0637,
+      "step": 3200
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.512370670265408e-06,
+      "loss": 0.2084,
+      "step": 3210
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.510571300044985e-06,
+      "loss": 0.1303,
+      "step": 3220
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.508771929824562e-06,
+      "loss": 0.1449,
+      "step": 3230
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.50697255960414e-06,
+      "loss": 0.0977,
+      "step": 3240
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.505173189383717e-06,
+      "loss": 0.0754,
+      "step": 3250
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.503373819163294e-06,
+      "loss": 0.0237,
+      "step": 3260
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.50157444894287e-06,
+      "loss": 0.1629,
+      "step": 3270
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.499775078722448e-06,
+      "loss": 0.1183,
+      "step": 3280
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.497975708502024e-06,
+      "loss": 0.0365,
+      "step": 3290
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.496176338281601e-06,
+      "loss": 0.0068,
+      "step": 3300
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.49437696806118e-06,
+      "loss": 0.2428,
+      "step": 3310
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.492577597840757e-06,
+      "loss": 0.0858,
+      "step": 3320
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.490778227620334e-06,
+      "loss": 0.0579,
+      "step": 3330
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.48897885739991e-06,
+      "loss": 0.0911,
+      "step": 3340
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.487179487179487e-06,
+      "loss": 0.0851,
+      "step": 3350
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.485380116959066e-06,
+      "loss": 0.1322,
+      "step": 3360
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.483580746738643e-06,
+      "loss": 0.0657,
+      "step": 3370
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.48178137651822e-06,
+      "loss": 0.0299,
+      "step": 3380
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.479982006297796e-06,
+      "loss": 0.2368,
+      "step": 3390
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.478182636077373e-06,
+      "loss": 0.064,
+      "step": 3400
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.47638326585695e-06,
+      "loss": 0.1288,
+      "step": 3410
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.474583895636527e-06,
+      "loss": 0.0765,
+      "step": 3420
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.472784525416106e-06,
+      "loss": 0.1403,
+      "step": 3430
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.470985155195682e-06,
+      "loss": 0.0662,
+      "step": 3440
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.46918578497526e-06,
+      "loss": 0.1429,
+      "step": 3450
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.467386414754836e-06,
+      "loss": 0.1111,
+      "step": 3460
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.465587044534413e-06,
+      "loss": 0.1807,
+      "step": 3470
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.463787674313992e-06,
+      "loss": 0.1044,
+      "step": 3480
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.461988304093568e-06,
+      "loss": 0.0654,
+      "step": 3490
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.460188933873145e-06,
+      "loss": 0.0255,
+      "step": 3500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.458389563652722e-06,
+      "loss": 0.1917,
+      "step": 3510
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.456590193432299e-06,
+      "loss": 0.02,
+      "step": 3520
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.454790823211878e-06,
+      "loss": 0.0694,
+      "step": 3530
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.452991452991453e-06,
+      "loss": 0.1769,
+      "step": 3540
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.451192082771031e-06,
+      "loss": 0.0723,
+      "step": 3550
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.449392712550608e-06,
+      "loss": 0.1041,
+      "step": 3560
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.447593342330185e-06,
+      "loss": 0.0711,
+      "step": 3570
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.445793972109762e-06,
+      "loss": 0.0932,
+      "step": 3580
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.443994601889339e-06,
+      "loss": 0.1224,
+      "step": 3590
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.442195231668917e-06,
+      "loss": 0.0452,
+      "step": 3600
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.440395861448494e-06,
+      "loss": 0.0763,
+      "step": 3610
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.438596491228071e-06,
+      "loss": 0.0422,
+      "step": 3620
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.436797121007648e-06,
+      "loss": 0.0184,
+      "step": 3630
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.434997750787225e-06,
+      "loss": 0.0059,
+      "step": 3640
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.433198380566803e-06,
+      "loss": 0.1277,
+      "step": 3650
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.431399010346378e-06,
+      "loss": 0.0845,
+      "step": 3660
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.429599640125957e-06,
+      "loss": 0.1707,
+      "step": 3670
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.427800269905534e-06,
+      "loss": 0.1214,
+      "step": 3680
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.42600089968511e-06,
+      "loss": 0.0493,
+      "step": 3690
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.424201529464688e-06,
+      "loss": 0.1305,
+      "step": 3700
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.422402159244264e-06,
+      "loss": 0.0809,
+      "step": 3710
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.420602789023843e-06,
+      "loss": 0.0707,
+      "step": 3720
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.41880341880342e-06,
+      "loss": 0.1068,
+      "step": 3730
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.417004048582997e-06,
+      "loss": 0.0152,
+      "step": 3740
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.415204678362574e-06,
+      "loss": 0.0773,
+      "step": 3750
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.41340530814215e-06,
+      "loss": 0.072,
+      "step": 3760
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.411605937921729e-06,
+      "loss": 0.1304,
+      "step": 3770
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.409806567701306e-06,
+      "loss": 0.0825,
+      "step": 3780
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.408007197480883e-06,
+      "loss": 0.1158,
+      "step": 3790
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.40620782726046e-06,
+      "loss": 0.2465,
+      "step": 3800
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.404408457040036e-06,
+      "loss": 0.2057,
+      "step": 3810
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.402609086819613e-06,
+      "loss": 0.1457,
+      "step": 3820
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.40080971659919e-06,
+      "loss": 0.0986,
+      "step": 3830
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.399010346378769e-06,
+      "loss": 0.1613,
+      "step": 3840
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.397210976158346e-06,
+      "loss": 0.0951,
+      "step": 3850
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.395411605937922e-06,
+      "loss": 0.094,
+      "step": 3860
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.3936122357175e-06,
+      "loss": 0.0467,
+      "step": 3870
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.391812865497076e-06,
+      "loss": 0.1046,
+      "step": 3880
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.390013495276655e-06,
+      "loss": 0.0671,
+      "step": 3890
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.388214125056232e-06,
+      "loss": 0.0632,
+      "step": 3900
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.386414754835808e-06,
+      "loss": 0.1289,
+      "step": 3910
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.384615384615385e-06,
+      "loss": 0.0962,
+      "step": 3920
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.382816014394962e-06,
+      "loss": 0.1108,
+      "step": 3930
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.381016644174539e-06,
+      "loss": 0.1097,
+      "step": 3940
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.379217273954116e-06,
+      "loss": 0.0337,
+      "step": 3950
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.377417903733694e-06,
+      "loss": 0.0564,
+      "step": 3960
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.375618533513271e-06,
+      "loss": 0.086,
+      "step": 3970
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.373819163292848e-06,
+      "loss": 0.0716,
+      "step": 3980
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.372019793072425e-06,
+      "loss": 0.0274,
+      "step": 3990
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.370220422852002e-06,
+      "loss": 0.0945,
+      "step": 4000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.36842105263158e-06,
+      "loss": 0.2289,
+      "step": 4010
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.366621682411157e-06,
+      "loss": 0.1001,
+      "step": 4020
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.364822312190734e-06,
+      "loss": 0.127,
+      "step": 4030
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.363022941970311e-06,
+      "loss": 0.0575,
+      "step": 4040
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.361223571749888e-06,
+      "loss": 0.14,
+      "step": 4050
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.359424201529466e-06,
+      "loss": 0.1394,
+      "step": 4060
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.357624831309042e-06,
+      "loss": 0.0982,
+      "step": 4070
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.35582546108862e-06,
+      "loss": 0.137,
+      "step": 4080
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.354026090868197e-06,
+      "loss": 0.0678,
+      "step": 4090
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.352226720647774e-06,
+      "loss": 0.1493,
+      "step": 4100
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.35042735042735e-06,
+      "loss": 0.0452,
+      "step": 4110
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.348627980206928e-06,
+      "loss": 0.0618,
+      "step": 4120
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.346828609986506e-06,
+      "loss": 0.0688,
+      "step": 4130
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.345029239766083e-06,
+      "loss": 0.0446,
+      "step": 4140
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.34322986954566e-06,
+      "loss": 0.1038,
+      "step": 4150
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.341430499325237e-06,
+      "loss": 0.0517,
+      "step": 4160
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.339631129104814e-06,
+      "loss": 0.1752,
+      "step": 4170
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.337831758884392e-06,
+      "loss": 0.0554,
+      "step": 4180
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.336032388663967e-06,
+      "loss": 0.0795,
+      "step": 4190
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.334233018443546e-06,
+      "loss": 0.0729,
+      "step": 4200
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.332433648223123e-06,
+      "loss": 0.1994,
+      "step": 4210
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.3306342780027e-06,
+      "loss": 0.1256,
+      "step": 4220
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.328834907782276e-06,
+      "loss": 0.0416,
+      "step": 4230
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.327035537561853e-06,
+      "loss": 0.0455,
+      "step": 4240
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.325236167341432e-06,
+      "loss": 0.1276,
+      "step": 4250
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.323436797121009e-06,
+      "loss": 0.1538,
+      "step": 4260
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.321637426900586e-06,
+      "loss": 0.072,
+      "step": 4270
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.319838056680162e-06,
+      "loss": 0.1352,
+      "step": 4280
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.31803868645974e-06,
+      "loss": 0.0613,
+      "step": 4290
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.316239316239318e-06,
+      "loss": 0.1611,
+      "step": 4300
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.314439946018893e-06,
+      "loss": 0.0851,
+      "step": 4310
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.312640575798472e-06,
+      "loss": 0.1095,
+      "step": 4320
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.310841205578048e-06,
+      "loss": 0.1378,
+      "step": 4330
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.309041835357625e-06,
+      "loss": 0.0981,
+      "step": 4340
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.307242465137202e-06,
+      "loss": 0.0396,
+      "step": 4350
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.305443094916779e-06,
+      "loss": 0.1012,
+      "step": 4360
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.303643724696358e-06,
+      "loss": 0.1195,
+      "step": 4370
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.301844354475934e-06,
+      "loss": 0.0449,
+      "step": 4380
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.300044984255511e-06,
+      "loss": 0.0778,
+      "step": 4390
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.298245614035088e-06,
+      "loss": 0.0422,
+      "step": 4400
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.296446243814665e-06,
+      "loss": 0.1016,
+      "step": 4410
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.294646873594244e-06,
+      "loss": 0.1021,
+      "step": 4420
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.29284750337382e-06,
+      "loss": 0.1569,
+      "step": 4430
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.291048133153397e-06,
+      "loss": 0.1313,
+      "step": 4440
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.289248762932974e-06,
+      "loss": 0.0682,
+      "step": 4450
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.287449392712551e-06,
+      "loss": 0.1345,
+      "step": 4460
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.285650022492128e-06,
+      "loss": 0.0751,
+      "step": 4470
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.283850652271705e-06,
+      "loss": 0.1235,
+      "step": 4480
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.282051282051283e-06,
+      "loss": 0.1123,
+      "step": 4490
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.28025191183086e-06,
+      "loss": 0.0496,
+      "step": 4500
+    },
+    {
+      "epoch": 0.4,
+      "eval_accuracy": 0.9711984841307437,
+      "eval_f1": 0.9508832399827659,
+      "eval_loss": 0.08512861281633377,
+      "eval_precision": 0.9297525013164823,
+      "eval_recall": 0.9729968037032954,
+      "eval_runtime": 436.7164,
+      "eval_samples_per_second": 72.507,
+      "eval_steps_per_second": 4.534,
+      "step": 4500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.278452541610437e-06,
+      "loss": 0.1278,
+      "step": 4510
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.276653171390014e-06,
+      "loss": 0.0659,
+      "step": 4520
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.27485380116959e-06,
+      "loss": 0.1262,
+      "step": 4530
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.27305443094917e-06,
+      "loss": 0.0633,
+      "step": 4540
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.271255060728746e-06,
+      "loss": 0.1094,
+      "step": 4550
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.269455690508323e-06,
+      "loss": 0.1054,
+      "step": 4560
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.2676563202879e-06,
+      "loss": 0.0375,
+      "step": 4570
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.265856950067477e-06,
+      "loss": 0.0434,
+      "step": 4580
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.264057579847054e-06,
+      "loss": 0.0875,
+      "step": 4590
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.26225820962663e-06,
+      "loss": 0.0963,
+      "step": 4600
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.260458839406209e-06,
+      "loss": 0.1243,
+      "step": 4610
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.258659469185786e-06,
+      "loss": 0.1053,
+      "step": 4620
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.256860098965363e-06,
+      "loss": 0.0566,
+      "step": 4630
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.25506072874494e-06,
+      "loss": 0.0947,
+      "step": 4640
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.253261358524516e-06,
+      "loss": 0.1326,
+      "step": 4650
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.251461988304095e-06,
+      "loss": 0.1622,
+      "step": 4660
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.249662618083672e-06,
+      "loss": 0.0438,
+      "step": 4670
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.247863247863249e-06,
+      "loss": 0.0346,
+      "step": 4680
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.246063877642826e-06,
+      "loss": 0.0647,
+      "step": 4690
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.244264507422402e-06,
+      "loss": 0.2105,
+      "step": 4700
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.242465137201981e-06,
+      "loss": 0.0517,
+      "step": 4710
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.240665766981556e-06,
+      "loss": 0.0369,
+      "step": 4720
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.238866396761135e-06,
+      "loss": 0.0792,
+      "step": 4730
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.237067026540712e-06,
+      "loss": 0.1478,
+      "step": 4740
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.235267656320288e-06,
+      "loss": 0.0625,
+      "step": 4750
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.233468286099865e-06,
+      "loss": 0.1502,
+      "step": 4760
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.231668915879442e-06,
+      "loss": 0.1159,
+      "step": 4770
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.22986954565902e-06,
+      "loss": 0.0383,
+      "step": 4780
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.228070175438598e-06,
+      "loss": 0.0419,
+      "step": 4790
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.226270805218174e-06,
+      "loss": 0.0887,
+      "step": 4800
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.224471434997751e-06,
+      "loss": 0.0085,
+      "step": 4810
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.222672064777328e-06,
+      "loss": 0.1119,
+      "step": 4820
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.220872694556907e-06,
+      "loss": 0.0901,
+      "step": 4830
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.219073324336482e-06,
+      "loss": 0.1066,
+      "step": 4840
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.21727395411606e-06,
+      "loss": 0.096,
+      "step": 4850
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.215474583895637e-06,
+      "loss": 0.1194,
+      "step": 4860
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.213675213675214e-06,
+      "loss": 0.0887,
+      "step": 4870
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.211875843454791e-06,
+      "loss": 0.0658,
+      "step": 4880
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.210076473234368e-06,
+      "loss": 0.1369,
+      "step": 4890
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.208277103013946e-06,
+      "loss": 0.1153,
+      "step": 4900
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.206477732793523e-06,
+      "loss": 0.0603,
+      "step": 4910
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.2046783625731e-06,
+      "loss": 0.1061,
+      "step": 4920
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.202878992352677e-06,
+      "loss": 0.0689,
+      "step": 4930
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.201079622132254e-06,
+      "loss": 0.1351,
+      "step": 4940
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.199280251911832e-06,
+      "loss": 0.1209,
+      "step": 4950
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.19748088169141e-06,
+      "loss": 0.0371,
+      "step": 4960
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.195681511470986e-06,
+      "loss": 0.0833,
+      "step": 4970
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.193882141250563e-06,
+      "loss": 0.1354,
+      "step": 4980
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.19208277103014e-06,
+      "loss": 0.2174,
+      "step": 4990
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.190283400809717e-06,
+      "loss": 0.1419,
+      "step": 5000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.188484030589294e-06,
+      "loss": 0.1198,
+      "step": 5010
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.186684660368872e-06,
+      "loss": 0.1137,
+      "step": 5020
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.184885290148449e-06,
+      "loss": 0.094,
+      "step": 5030
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.183085919928026e-06,
+      "loss": 0.0805,
+      "step": 5040
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.181286549707603e-06,
+      "loss": 0.0516,
+      "step": 5050
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.17948717948718e-06,
+      "loss": 0.0696,
+      "step": 5060
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.177687809266758e-06,
+      "loss": 0.0769,
+      "step": 5070
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.175888439046335e-06,
+      "loss": 0.0959,
+      "step": 5080
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.174089068825912e-06,
+      "loss": 0.0542,
+      "step": 5090
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.172289698605489e-06,
+      "loss": 0.0712,
+      "step": 5100
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.170490328385066e-06,
+      "loss": 0.0537,
+      "step": 5110
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.168690958164642e-06,
+      "loss": 0.1163,
+      "step": 5120
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.16689158794422e-06,
+      "loss": 0.0765,
+      "step": 5130
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.165092217723798e-06,
+      "loss": 0.2003,
+      "step": 5140
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.163292847503375e-06,
+      "loss": 0.0108,
+      "step": 5150
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.161493477282952e-06,
+      "loss": 0.1257,
+      "step": 5160
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.159694107062528e-06,
+      "loss": 0.1449,
+      "step": 5170
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.157894736842105e-06,
+      "loss": 0.053,
+      "step": 5180
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.156095366621684e-06,
+      "loss": 0.0475,
+      "step": 5190
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.15429599640126e-06,
+      "loss": 0.029,
+      "step": 5200
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.152496626180838e-06,
+      "loss": 0.0036,
+      "step": 5210
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.150697255960414e-06,
+      "loss": 0.0076,
+      "step": 5220
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.148897885739991e-06,
+      "loss": 0.0655,
+      "step": 5230
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.14709851551957e-06,
+      "loss": 0.0534,
+      "step": 5240
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.145299145299145e-06,
+      "loss": 0.0378,
+      "step": 5250
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.143499775078724e-06,
+      "loss": 0.1112,
+      "step": 5260
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.1417004048583e-06,
+      "loss": 0.1258,
+      "step": 5270
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.139901034637877e-06,
+      "loss": 0.1018,
+      "step": 5280
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.138101664417454e-06,
+      "loss": 0.1015,
+      "step": 5290
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.136302294197031e-06,
+      "loss": 0.2089,
+      "step": 5300
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.13450292397661e-06,
+      "loss": 0.0352,
+      "step": 5310
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.132703553756186e-06,
+      "loss": 0.1233,
+      "step": 5320
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.130904183535763e-06,
+      "loss": 0.1092,
+      "step": 5330
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.12910481331534e-06,
+      "loss": 0.0103,
+      "step": 5340
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.127305443094917e-06,
+      "loss": 0.0304,
+      "step": 5350
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.125506072874496e-06,
+      "loss": 0.1365,
+      "step": 5360
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.12370670265407e-06,
+      "loss": 0.1332,
+      "step": 5370
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.12190733243365e-06,
+      "loss": 0.0738,
+      "step": 5380
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.120107962213226e-06,
+      "loss": 0.1559,
+      "step": 5390
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.118308591992803e-06,
+      "loss": 0.139,
+      "step": 5400
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.11650922177238e-06,
+      "loss": 0.1414,
+      "step": 5410
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.114709851551957e-06,
+      "loss": 0.1316,
+      "step": 5420
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.112910481331535e-06,
+      "loss": 0.1166,
+      "step": 5430
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.111111111111112e-06,
+      "loss": 0.0909,
+      "step": 5440
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.109311740890689e-06,
+      "loss": 0.1103,
+      "step": 5450
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.107512370670266e-06,
+      "loss": 0.1618,
+      "step": 5460
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.105713000449843e-06,
+      "loss": 0.0721,
+      "step": 5470
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.103913630229421e-06,
+      "loss": 0.1584,
+      "step": 5480
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.102114260008998e-06,
+      "loss": 0.1035,
+      "step": 5490
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.100314889788575e-06,
+      "loss": 0.028,
+      "step": 5500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.098515519568152e-06,
+      "loss": 0.1161,
+      "step": 5510
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.096716149347729e-06,
+      "loss": 0.1424,
+      "step": 5520
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.094916779127306e-06,
+      "loss": 0.1636,
+      "step": 5530
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.093117408906882e-06,
+      "loss": 0.1167,
+      "step": 5540
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.091318038686461e-06,
+      "loss": 0.0831,
+      "step": 5550
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.089518668466038e-06,
+      "loss": 0.1418,
+      "step": 5560
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.087719298245615e-06,
+      "loss": 0.1449,
+      "step": 5570
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.085919928025192e-06,
+      "loss": 0.1729,
+      "step": 5580
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.084120557804768e-06,
+      "loss": 0.125,
+      "step": 5590
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.082321187584347e-06,
+      "loss": 0.0741,
+      "step": 5600
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.080521817363924e-06,
+      "loss": 0.0811,
+      "step": 5610
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.0787224471435e-06,
+      "loss": 0.1593,
+      "step": 5620
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.076923076923078e-06,
+      "loss": 0.1943,
+      "step": 5630
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.075123706702654e-06,
+      "loss": 0.0983,
+      "step": 5640
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.073324336482231e-06,
+      "loss": 0.1401,
+      "step": 5650
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.071524966261808e-06,
+      "loss": 0.1739,
+      "step": 5660
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.069725596041387e-06,
+      "loss": 0.1541,
+      "step": 5670
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.067926225820964e-06,
+      "loss": 0.1699,
+      "step": 5680
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.06612685560054e-06,
+      "loss": 0.0351,
+      "step": 5690
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.064327485380117e-06,
+      "loss": 0.0941,
+      "step": 5700
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.062528115159694e-06,
+      "loss": 0.1932,
+      "step": 5710
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.060728744939273e-06,
+      "loss": 0.0819,
+      "step": 5720
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.05892937471885e-06,
+      "loss": 0.2269,
+      "step": 5730
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.057130004498426e-06,
+      "loss": 0.0967,
+      "step": 5740
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.055330634278003e-06,
+      "loss": 0.0594,
+      "step": 5750
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.05353126405758e-06,
+      "loss": 0.2919,
+      "step": 5760
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.051731893837159e-06,
+      "loss": 0.0618,
+      "step": 5770
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.049932523616734e-06,
+      "loss": 0.1166,
+      "step": 5780
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.048133153396312e-06,
+      "loss": 0.124,
+      "step": 5790
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.04633378317589e-06,
+      "loss": 0.1845,
+      "step": 5800
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.044534412955466e-06,
+      "loss": 0.1357,
+      "step": 5810
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.042735042735043e-06,
+      "loss": 0.1428,
+      "step": 5820
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.04093567251462e-06,
+      "loss": 0.1932,
+      "step": 5830
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.039136302294198e-06,
+      "loss": 0.0585,
+      "step": 5840
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.037336932073775e-06,
+      "loss": 0.0983,
+      "step": 5850
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.035537561853352e-06,
+      "loss": 0.0764,
+      "step": 5860
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.033738191632929e-06,
+      "loss": 0.1484,
+      "step": 5870
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.031938821412506e-06,
+      "loss": 0.1209,
+      "step": 5880
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.030139451192084e-06,
+      "loss": 0.0421,
+      "step": 5890
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.02834008097166e-06,
+      "loss": 0.0788,
+      "step": 5900
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.026540710751238e-06,
+      "loss": 0.0637,
+      "step": 5910
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.024741340530815e-06,
+      "loss": 0.0962,
+      "step": 5920
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.022941970310392e-06,
+      "loss": 0.0681,
+      "step": 5930
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.021142600089969e-06,
+      "loss": 0.1625,
+      "step": 5940
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.019343229869546e-06,
+      "loss": 0.1218,
+      "step": 5950
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.017543859649124e-06,
+      "loss": 0.0072,
+      "step": 5960
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.015744489428701e-06,
+      "loss": 0.1611,
+      "step": 5970
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.013945119208278e-06,
+      "loss": 0.1108,
+      "step": 5980
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.012145748987855e-06,
+      "loss": 0.1706,
+      "step": 5990
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.010346378767432e-06,
+      "loss": 0.0967,
+      "step": 6000
+    },
+    {
+      "epoch": 0.53,
+      "eval_accuracy": 0.9735354492341702,
+      "eval_f1": 0.954082191780822,
+      "eval_loss": 0.09180905669927597,
+      "eval_precision": 0.9486760379208892,
+      "eval_recall": 0.9595503141188141,
+      "eval_runtime": 436.4287,
+      "eval_samples_per_second": 72.555,
+      "eval_steps_per_second": 4.537,
+      "step": 6000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.00854700854701e-06,
+      "loss": 0.0714,
+      "step": 6010
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.006747638326587e-06,
+      "loss": 0.1627,
+      "step": 6020
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.004948268106164e-06,
+      "loss": 0.0399,
+      "step": 6030
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.00314889788574e-06,
+      "loss": 0.0568,
+      "step": 6040
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.001349527665318e-06,
+      "loss": 0.1471,
+      "step": 6050
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.999550157444894e-06,
+      "loss": 0.0979,
+      "step": 6060
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.997750787224471e-06,
+      "loss": 0.0177,
+      "step": 6070
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.99595141700405e-06,
+      "loss": 0.1574,
+      "step": 6080
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.994152046783627e-06,
+      "loss": 0.1683,
+      "step": 6090
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.992352676563204e-06,
+      "loss": 0.1035,
+      "step": 6100
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.99055330634278e-06,
+      "loss": 0.1381,
+      "step": 6110
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.988753936122357e-06,
+      "loss": 0.1026,
+      "step": 6120
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.986954565901936e-06,
+      "loss": 0.0454,
+      "step": 6130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.985155195681513e-06,
+      "loss": 0.0653,
+      "step": 6140
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.98335582546109e-06,
+      "loss": 0.0681,
+      "step": 6150
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.981556455240666e-06,
+      "loss": 0.1286,
+      "step": 6160
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.979757085020243e-06,
+      "loss": 0.0823,
+      "step": 6170
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.97795771479982e-06,
+      "loss": 0.0489,
+      "step": 6180
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.976158344579397e-06,
+      "loss": 0.0605,
+      "step": 6190
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.974358974358976e-06,
+      "loss": 0.054,
+      "step": 6200
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.972559604138552e-06,
+      "loss": 0.1616,
+      "step": 6210
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.97076023391813e-06,
+      "loss": 0.0963,
+      "step": 6220
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.968960863697706e-06,
+      "loss": 0.1245,
+      "step": 6230
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.967161493477283e-06,
+      "loss": 0.092,
+      "step": 6240
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.965362123256862e-06,
+      "loss": 0.0728,
+      "step": 6250
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.963562753036438e-06,
+      "loss": 0.1138,
+      "step": 6260
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.961763382816015e-06,
+      "loss": 0.1143,
+      "step": 6270
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.959964012595592e-06,
+      "loss": 0.0954,
+      "step": 6280
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.958164642375169e-06,
+      "loss": 0.0606,
+      "step": 6290
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.956365272154746e-06,
+      "loss": 0.0667,
+      "step": 6300
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.954565901934323e-06,
+      "loss": 0.0505,
+      "step": 6310
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.952766531713901e-06,
+      "loss": 0.1699,
+      "step": 6320
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.950967161493478e-06,
+      "loss": 0.0568,
+      "step": 6330
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.949167791273055e-06,
+      "loss": 0.1413,
+      "step": 6340
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.947368421052632e-06,
+      "loss": 0.0903,
+      "step": 6350
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.945569050832209e-06,
+      "loss": 0.1166,
+      "step": 6360
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.943769680611787e-06,
+      "loss": 0.0781,
+      "step": 6370
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.941970310391364e-06,
+      "loss": 0.0385,
+      "step": 6380
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.940170940170941e-06,
+      "loss": 0.0508,
+      "step": 6390
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.938371569950518e-06,
+      "loss": 0.0893,
+      "step": 6400
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.936572199730095e-06,
+      "loss": 0.1399,
+      "step": 6410
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.934772829509673e-06,
+      "loss": 0.0252,
+      "step": 6420
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.932973459289248e-06,
+      "loss": 0.0843,
+      "step": 6430
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.931174089068827e-06,
+      "loss": 0.0468,
+      "step": 6440
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.929374718848404e-06,
+      "loss": 0.0843,
+      "step": 6450
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.92757534862798e-06,
+      "loss": 0.0478,
+      "step": 6460
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.925775978407558e-06,
+      "loss": 0.1855,
+      "step": 6470
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.923976608187134e-06,
+      "loss": 0.033,
+      "step": 6480
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.922177237966713e-06,
+      "loss": 0.1028,
+      "step": 6490
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.92037786774629e-06,
+      "loss": 0.1134,
+      "step": 6500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.918578497525867e-06,
+      "loss": 0.0714,
+      "step": 6510
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.916779127305444e-06,
+      "loss": 0.1058,
+      "step": 6520
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.91497975708502e-06,
+      "loss": 0.0672,
+      "step": 6530
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.913180386864599e-06,
+      "loss": 0.1081,
+      "step": 6540
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.911381016644174e-06,
+      "loss": 0.0558,
+      "step": 6550
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.909581646423753e-06,
+      "loss": 0.1227,
+      "step": 6560
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.90778227620333e-06,
+      "loss": 0.1211,
+      "step": 6570
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.905982905982906e-06,
+      "loss": 0.0534,
+      "step": 6580
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.904183535762483e-06,
+      "loss": 0.1179,
+      "step": 6590
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.90238416554206e-06,
+      "loss": 0.1224,
+      "step": 6600
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.900584795321639e-06,
+      "loss": 0.0878,
+      "step": 6610
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.898785425101216e-06,
+      "loss": 0.12,
+      "step": 6620
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.896986054880792e-06,
+      "loss": 0.1773,
+      "step": 6630
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.89518668466037e-06,
+      "loss": 0.0991,
+      "step": 6640
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.893387314439946e-06,
+      "loss": 0.1262,
+      "step": 6650
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.891587944219525e-06,
+      "loss": 0.0043,
+      "step": 6660
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 8.889788573999102e-06,
+      "loss": 0.0821,
+      "step": 6670
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.887989203778678e-06,
+      "loss": 0.212,
+      "step": 6680
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.886189833558255e-06,
+      "loss": 0.0983,
+      "step": 6690
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.884390463337832e-06,
+      "loss": 0.1402,
+      "step": 6700
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.882591093117409e-06,
+      "loss": 0.159,
+      "step": 6710
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.880791722896986e-06,
+      "loss": 0.1258,
+      "step": 6720
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.878992352676564e-06,
+      "loss": 0.0954,
+      "step": 6730
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.877192982456141e-06,
+      "loss": 0.0412,
+      "step": 6740
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.875393612235718e-06,
+      "loss": 0.078,
+      "step": 6750
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.873594242015295e-06,
+      "loss": 0.1009,
+      "step": 6760
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.871794871794872e-06,
+      "loss": 0.099,
+      "step": 6770
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 8.86999550157445e-06,
+      "loss": 0.0313,
+      "step": 6780
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.868196131354027e-06,
+      "loss": 0.0375,
+      "step": 6790
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.866396761133604e-06,
+      "loss": 0.0885,
+      "step": 6800
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.864597390913181e-06,
+      "loss": 0.0761,
+      "step": 6810
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.862798020692758e-06,
+      "loss": 0.0892,
+      "step": 6820
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.860998650472335e-06,
+      "loss": 0.1141,
+      "step": 6830
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.859199280251912e-06,
+      "loss": 0.0784,
+      "step": 6840
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.85739991003149e-06,
+      "loss": 0.1206,
+      "step": 6850
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.855600539811067e-06,
+      "loss": 0.054,
+      "step": 6860
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.853801169590644e-06,
+      "loss": 0.2208,
+      "step": 6870
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.85200179937022e-06,
+      "loss": 0.0689,
+      "step": 6880
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 8.850202429149798e-06,
+      "loss": 0.0898,
+      "step": 6890
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.848403058929376e-06,
+      "loss": 0.1601,
+      "step": 6900
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.846603688708953e-06,
+      "loss": 0.0574,
+      "step": 6910
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.84480431848853e-06,
+      "loss": 0.0564,
+      "step": 6920
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.843004948268107e-06,
+      "loss": 0.0072,
+      "step": 6930
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.841205578047684e-06,
+      "loss": 0.0368,
+      "step": 6940
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.839406207827262e-06,
+      "loss": 0.1152,
+      "step": 6950
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.837606837606837e-06,
+      "loss": 0.0871,
+      "step": 6960
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.835807467386416e-06,
+      "loss": 0.0936,
+      "step": 6970
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.834008097165993e-06,
+      "loss": 0.2966,
+      "step": 6980
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.83220872694557e-06,
+      "loss": 0.1432,
+      "step": 6990
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 8.830409356725146e-06,
+      "loss": 0.135,
+      "step": 7000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.828609986504723e-06,
+      "loss": 0.0894,
+      "step": 7010
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.826810616284302e-06,
+      "loss": 0.1052,
+      "step": 7020
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.825011246063879e-06,
+      "loss": 0.1165,
+      "step": 7030
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.823211875843456e-06,
+      "loss": 0.0955,
+      "step": 7040
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.821412505623032e-06,
+      "loss": 0.0356,
+      "step": 7050
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.81961313540261e-06,
+      "loss": 0.143,
+      "step": 7060
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.817813765182188e-06,
+      "loss": 0.0743,
+      "step": 7070
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.816014394961763e-06,
+      "loss": 0.0865,
+      "step": 7080
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.814215024741342e-06,
+      "loss": 0.1262,
+      "step": 7090
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.812415654520918e-06,
+      "loss": 0.0891,
+      "step": 7100
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.810616284300495e-06,
+      "loss": 0.1062,
+      "step": 7110
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 8.808816914080072e-06,
+      "loss": 0.0651,
+      "step": 7120
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.807017543859649e-06,
+      "loss": 0.1016,
+      "step": 7130
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.805218173639228e-06,
+      "loss": 0.1007,
+      "step": 7140
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.803418803418804e-06,
+      "loss": 0.0866,
+      "step": 7150
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.801619433198381e-06,
+      "loss": 0.0866,
+      "step": 7160
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.799820062977958e-06,
+      "loss": 0.0524,
+      "step": 7170
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.798020692757535e-06,
+      "loss": 0.0092,
+      "step": 7180
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.796221322537114e-06,
+      "loss": 0.0843,
+      "step": 7190
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.79442195231669e-06,
+      "loss": 0.0962,
+      "step": 7200
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.792622582096267e-06,
+      "loss": 0.0755,
+      "step": 7210
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.790823211875844e-06,
+      "loss": 0.1832,
+      "step": 7220
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.789023841655421e-06,
+      "loss": 0.1657,
+      "step": 7230
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.787224471434998e-06,
+      "loss": 0.0969,
+      "step": 7240
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.785425101214575e-06,
+      "loss": 0.1005,
+      "step": 7250
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.783625730994153e-06,
+      "loss": 0.0896,
+      "step": 7260
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.78182636077373e-06,
+      "loss": 0.0941,
+      "step": 7270
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.780026990553307e-06,
+      "loss": 0.1214,
+      "step": 7280
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.778227620332884e-06,
+      "loss": 0.1319,
+      "step": 7290
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.77642825011246e-06,
+      "loss": 0.0806,
+      "step": 7300
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.77462887989204e-06,
+      "loss": 0.1304,
+      "step": 7310
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.772829509671616e-06,
+      "loss": 0.0723,
+      "step": 7320
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.771030139451193e-06,
+      "loss": 0.176,
+      "step": 7330
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.76923076923077e-06,
+      "loss": 0.0992,
+      "step": 7340
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.767431399010347e-06,
+      "loss": 0.128,
+      "step": 7350
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.765632028789924e-06,
+      "loss": 0.0585,
+      "step": 7360
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.7638326585695e-06,
+      "loss": 0.2104,
+      "step": 7370
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.762033288349079e-06,
+      "loss": 0.0491,
+      "step": 7380
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.760233918128656e-06,
+      "loss": 0.1178,
+      "step": 7390
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.758434547908233e-06,
+      "loss": 0.181,
+      "step": 7400
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.75663517768781e-06,
+      "loss": 0.1329,
+      "step": 7410
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.754835807467386e-06,
+      "loss": 0.1067,
+      "step": 7420
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.753036437246965e-06,
+      "loss": 0.1344,
+      "step": 7430
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.751237067026542e-06,
+      "loss": 0.0359,
+      "step": 7440
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.749437696806119e-06,
+      "loss": 0.0587,
+      "step": 7450
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.747638326585696e-06,
+      "loss": 0.0763,
+      "step": 7460
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.745838956365272e-06,
+      "loss": 0.1267,
+      "step": 7470
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.744039586144851e-06,
+      "loss": 0.0407,
+      "step": 7480
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.742240215924426e-06,
+      "loss": 0.0741,
+      "step": 7490
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.740440845704005e-06,
+      "loss": 0.0914,
+      "step": 7500
+    },
+    {
+      "epoch": 0.67,
+      "eval_accuracy": 0.971293225959261,
+      "eval_f1": 0.95092057664273,
+      "eval_loss": 0.12410593777894974,
+      "eval_precision": 0.9320491109229466,
+      "eval_recall": 0.9705720268929792,
+      "eval_runtime": 436.6013,
+      "eval_samples_per_second": 72.526,
+      "eval_steps_per_second": 4.535,
+      "step": 7500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.738641475483582e-06,
+      "loss": 0.1633,
+      "step": 7510
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.736842105263158e-06,
+      "loss": 0.1843,
+      "step": 7520
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.735042735042735e-06,
+      "loss": 0.1008,
+      "step": 7530
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.733243364822312e-06,
+      "loss": 0.0962,
+      "step": 7540
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.73144399460189e-06,
+      "loss": 0.0271,
+      "step": 7550
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.729644624381468e-06,
+      "loss": 0.1435,
+      "step": 7560
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 8.727845254161044e-06,
+      "loss": 0.0794,
+      "step": 7570
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.726045883940621e-06,
+      "loss": 0.1544,
+      "step": 7580
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.724246513720198e-06,
+      "loss": 0.0104,
+      "step": 7590
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.722447143499777e-06,
+      "loss": 0.0687,
+      "step": 7600
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.720647773279352e-06,
+      "loss": 0.1329,
+      "step": 7610
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.71884840305893e-06,
+      "loss": 0.0721,
+      "step": 7620
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.717049032838507e-06,
+      "loss": 0.1454,
+      "step": 7630
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.715249662618084e-06,
+      "loss": 0.0498,
+      "step": 7640
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.713450292397661e-06,
+      "loss": 0.1081,
+      "step": 7650
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.711650922177238e-06,
+      "loss": 0.072,
+      "step": 7660
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.709851551956816e-06,
+      "loss": 0.0765,
+      "step": 7670
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 8.708052181736393e-06,
+      "loss": 0.1375,
+      "step": 7680
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.70625281151597e-06,
+      "loss": 0.1537,
+      "step": 7690
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.704453441295547e-06,
+      "loss": 0.097,
+      "step": 7700
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.702654071075124e-06,
+      "loss": 0.1346,
+      "step": 7710
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.700854700854702e-06,
+      "loss": 0.1619,
+      "step": 7720
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.69905533063428e-06,
+      "loss": 0.1027,
+      "step": 7730
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.697255960413856e-06,
+      "loss": 0.1377,
+      "step": 7740
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.695456590193433e-06,
+      "loss": 0.0791,
+      "step": 7750
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.693657219973012e-06,
+      "loss": 0.0814,
+      "step": 7760
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.691857849752587e-06,
+      "loss": 0.078,
+      "step": 7770
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.690058479532164e-06,
+      "loss": 0.0902,
+      "step": 7780
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 8.688259109311742e-06,
+      "loss": 0.144,
+      "step": 7790
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.686459739091319e-06,
+      "loss": 0.1267,
+      "step": 7800
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.684660368870896e-06,
+      "loss": 0.0564,
+      "step": 7810
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.682860998650473e-06,
+      "loss": 0.0782,
+      "step": 7820
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.68106162843005e-06,
+      "loss": 0.1353,
+      "step": 7830
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.679262258209628e-06,
+      "loss": 0.0428,
+      "step": 7840
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.677462887989205e-06,
+      "loss": 0.0864,
+      "step": 7850
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.675663517768782e-06,
+      "loss": 0.0717,
+      "step": 7860
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.673864147548359e-06,
+      "loss": 0.1192,
+      "step": 7870
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.672064777327936e-06,
+      "loss": 0.0728,
+      "step": 7880
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.670265407107512e-06,
+      "loss": 0.0615,
+      "step": 7890
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 8.66846603688709e-06,
+      "loss": 0.0884,
+      "step": 7900
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.666666666666668e-06,
+      "loss": 0.1488,
+      "step": 7910
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.664867296446245e-06,
+      "loss": 0.1203,
+      "step": 7920
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.663067926225822e-06,
+      "loss": 0.0469,
+      "step": 7930
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.661268556005398e-06,
+      "loss": 0.0475,
+      "step": 7940
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.659469185784975e-06,
+      "loss": 0.1008,
+      "step": 7950
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.657669815564554e-06,
+      "loss": 0.0544,
+      "step": 7960
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.65587044534413e-06,
+      "loss": 0.1543,
+      "step": 7970
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.654071075123708e-06,
+      "loss": 0.1327,
+      "step": 7980
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.652271704903284e-06,
+      "loss": 0.3129,
+      "step": 7990
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.650472334682861e-06,
+      "loss": 0.0302,
+      "step": 8000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.64867296446244e-06,
+      "loss": 0.1717,
+      "step": 8010
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.646873594242015e-06,
+      "loss": 0.0878,
+      "step": 8020
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.645074224021594e-06,
+      "loss": 0.0869,
+      "step": 8030
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.64327485380117e-06,
+      "loss": 0.0919,
+      "step": 8040
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.641475483580747e-06,
+      "loss": 0.058,
+      "step": 8050
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.639676113360324e-06,
+      "loss": 0.1208,
+      "step": 8060
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.637876743139901e-06,
+      "loss": 0.0707,
+      "step": 8070
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.63607737291948e-06,
+      "loss": 0.0532,
+      "step": 8080
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.634278002699056e-06,
+      "loss": 0.0889,
+      "step": 8090
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.632478632478633e-06,
+      "loss": 0.0517,
+      "step": 8100
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.63067926225821e-06,
+      "loss": 0.1343,
+      "step": 8110
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.628879892037787e-06,
+      "loss": 0.0304,
+      "step": 8120
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.627080521817366e-06,
+      "loss": 0.1531,
+      "step": 8130
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.62528115159694e-06,
+      "loss": 0.1384,
+      "step": 8140
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.62348178137652e-06,
+      "loss": 0.0572,
+      "step": 8150
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.621682411156096e-06,
+      "loss": 0.043,
+      "step": 8160
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.619883040935673e-06,
+      "loss": 0.0598,
+      "step": 8170
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.61808367071525e-06,
+      "loss": 0.0271,
+      "step": 8180
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.616284300494827e-06,
+      "loss": 0.0278,
+      "step": 8190
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.614484930274405e-06,
+      "loss": 0.0878,
+      "step": 8200
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.612685560053982e-06,
+      "loss": 0.1291,
+      "step": 8210
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.610886189833559e-06,
+      "loss": 0.091,
+      "step": 8220
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.609086819613136e-06,
+      "loss": 0.053,
+      "step": 8230
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.607287449392713e-06,
+      "loss": 0.0576,
+      "step": 8240
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.605488079172291e-06,
+      "loss": 0.1051,
+      "step": 8250
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.603688708951866e-06,
+      "loss": 0.0692,
+      "step": 8260
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.601889338731445e-06,
+      "loss": 0.1337,
+      "step": 8270
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.600089968511022e-06,
+      "loss": 0.0394,
+      "step": 8280
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.598290598290599e-06,
+      "loss": 0.066,
+      "step": 8290
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.596491228070176e-06,
+      "loss": 0.0414,
+      "step": 8300
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.594691857849752e-06,
+      "loss": 0.1076,
+      "step": 8310
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.592892487629331e-06,
+      "loss": 0.1802,
+      "step": 8320
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.591093117408908e-06,
+      "loss": 0.0185,
+      "step": 8330
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.589293747188485e-06,
+      "loss": 0.0604,
+      "step": 8340
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.587494376968062e-06,
+      "loss": 0.1542,
+      "step": 8350
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.585695006747638e-06,
+      "loss": 0.1218,
+      "step": 8360
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.583895636527217e-06,
+      "loss": 0.1441,
+      "step": 8370
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.582096266306794e-06,
+      "loss": 0.0921,
+      "step": 8380
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.58029689608637e-06,
+      "loss": 0.0408,
+      "step": 8390
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.578497525865948e-06,
+      "loss": 0.0247,
+      "step": 8400
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.576698155645524e-06,
+      "loss": 0.0663,
+      "step": 8410
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.574898785425101e-06,
+      "loss": 0.148,
+      "step": 8420
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.573099415204678e-06,
+      "loss": 0.0787,
+      "step": 8430
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.571300044984257e-06,
+      "loss": 0.0468,
+      "step": 8440
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.569500674763834e-06,
+      "loss": 0.1468,
+      "step": 8450
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.56770130454341e-06,
+      "loss": 0.0841,
+      "step": 8460
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.565901934322987e-06,
+      "loss": 0.0974,
+      "step": 8470
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.564102564102564e-06,
+      "loss": 0.0644,
+      "step": 8480
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.562303193882143e-06,
+      "loss": 0.055,
+      "step": 8490
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.56050382366172e-06,
+      "loss": 0.0476,
+      "step": 8500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.558704453441296e-06,
+      "loss": 0.2432,
+      "step": 8510
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.556905083220873e-06,
+      "loss": 0.1153,
+      "step": 8520
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.55510571300045e-06,
+      "loss": 0.0691,
+      "step": 8530
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.553306342780027e-06,
+      "loss": 0.0634,
+      "step": 8540
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.551506972559604e-06,
+      "loss": 0.1217,
+      "step": 8550
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.549707602339182e-06,
+      "loss": 0.15,
+      "step": 8560
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.54790823211876e-06,
+      "loss": 0.1146,
+      "step": 8570
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.546108861898336e-06,
+      "loss": 0.0232,
+      "step": 8580
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.544309491677913e-06,
+      "loss": 0.1652,
+      "step": 8590
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.54251012145749e-06,
+      "loss": 0.1147,
+      "step": 8600
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.540710751237068e-06,
+      "loss": 0.1156,
+      "step": 8610
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.538911381016645e-06,
+      "loss": 0.0725,
+      "step": 8620
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.537112010796222e-06,
+      "loss": 0.0058,
+      "step": 8630
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.535312640575799e-06,
+      "loss": 0.1656,
+      "step": 8640
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.533513270355376e-06,
+      "loss": 0.0576,
+      "step": 8650
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.531713900134954e-06,
+      "loss": 0.1474,
+      "step": 8660
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.52991452991453e-06,
+      "loss": 0.1511,
+      "step": 8670
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.528115159694108e-06,
+      "loss": 0.1008,
+      "step": 8680
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.526315789473685e-06,
+      "loss": 0.0811,
+      "step": 8690
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.524516419253262e-06,
+      "loss": 0.0772,
+      "step": 8700
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.522717049032839e-06,
+      "loss": 0.0955,
+      "step": 8710
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.520917678812416e-06,
+      "loss": 0.0719,
+      "step": 8720
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.519118308591994e-06,
+      "loss": 0.1383,
+      "step": 8730
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.517318938371571e-06,
+      "loss": 0.0035,
+      "step": 8740
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.515519568151148e-06,
+      "loss": 0.1452,
+      "step": 8750
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.513720197930725e-06,
+      "loss": 0.1435,
+      "step": 8760
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.511920827710302e-06,
+      "loss": 0.1142,
+      "step": 8770
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.51012145748988e-06,
+      "loss": 0.1441,
+      "step": 8780
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.508322087269455e-06,
+      "loss": 0.179,
+      "step": 8790
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.506522717049034e-06,
+      "loss": 0.0934,
+      "step": 8800
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.50472334682861e-06,
+      "loss": 0.0647,
+      "step": 8810
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.502923976608188e-06,
+      "loss": 0.0749,
+      "step": 8820
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.501124606387764e-06,
+      "loss": 0.1038,
+      "step": 8830
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.499325236167341e-06,
+      "loss": 0.0616,
+      "step": 8840
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.49752586594692e-06,
+      "loss": 0.1128,
+      "step": 8850
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.495726495726497e-06,
+      "loss": 0.1401,
+      "step": 8860
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.493927125506074e-06,
+      "loss": 0.2143,
+      "step": 8870
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.49212775528565e-06,
+      "loss": 0.0586,
+      "step": 8880
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.490328385065227e-06,
+      "loss": 0.0917,
+      "step": 8890
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.488529014844806e-06,
+      "loss": 0.1244,
+      "step": 8900
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.486729644624383e-06,
+      "loss": 0.0877,
+      "step": 8910
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.48493027440396e-06,
+      "loss": 0.0792,
+      "step": 8920
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.483130904183536e-06,
+      "loss": 0.1058,
+      "step": 8930
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.481331533963115e-06,
+      "loss": 0.1133,
+      "step": 8940
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.47953216374269e-06,
+      "loss": 0.0392,
+      "step": 8950
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.477732793522267e-06,
+      "loss": 0.1745,
+      "step": 8960
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.475933423301846e-06,
+      "loss": 0.1019,
+      "step": 8970
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.474134053081422e-06,
+      "loss": 0.166,
+      "step": 8980
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.472334682861e-06,
+      "loss": 0.1927,
+      "step": 8990
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.470535312640576e-06,
+      "loss": 0.1251,
+      "step": 9000
+    },
+    {
+      "epoch": 0.8,
+      "eval_accuracy": 0.9701563240170535,
+      "eval_f1": 0.9492399419885051,
+      "eval_loss": 0.09671162813901901,
+      "eval_precision": 0.9258172673931265,
+      "eval_recall": 0.9738785407252287,
+      "eval_runtime": 437.2054,
+      "eval_samples_per_second": 72.426,
+      "eval_steps_per_second": 4.529,
+      "step": 9000
+    }
+  ],
+  "max_steps": 56075,
+  "num_train_epochs": 5,
+  "total_flos": 7.6243123003392e+16,
+  "trial_name": null,
+  "trial_params": null
+}