diff --git "a/tmp-checkpoint-11707/trainer_state.json" "b/tmp-checkpoint-11707/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/tmp-checkpoint-11707/trainer_state.json"
@@ -0,0 +1,7461 @@
+{
+  "best_metric": 0.07501205056905746,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-crack-detectorMAIN50epochsFINAL\\checkpoint-11707",
+  "epoch": 20.999103139013453,
+  "eval_steps": 500,
+  "global_step": 11707,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.7953321364452426e-07,
+      "loss": 1.4257,
+      "step": 10
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.590664272890485e-07,
+      "loss": 1.4227,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.385996409335728e-07,
+      "loss": 1.4112,
+      "step": 30
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 7.18132854578097e-07,
+      "loss": 1.4052,
+      "step": 40
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 8.976660682226213e-07,
+      "loss": 1.3954,
+      "step": 50
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.0771992818671456e-06,
+      "loss": 1.3978,
+      "step": 60
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.2567324955116697e-06,
+      "loss": 1.3862,
+      "step": 70
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.436265709156194e-06,
+      "loss": 1.3651,
+      "step": 80
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.6157989228007182e-06,
+      "loss": 1.359,
+      "step": 90
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.7953321364452425e-06,
+      "loss": 1.3476,
+      "step": 100
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9748653500897667e-06,
+      "loss": 1.3303,
+      "step": 110
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.1543985637342912e-06,
+      "loss": 1.3047,
+      "step": 120
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.333931777378815e-06,
+      "loss": 1.2944,
+      "step": 130
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.5134649910233395e-06,
+      "loss": 1.284,
+      "step": 140
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.6929982046678636e-06,
+      "loss": 1.23,
+      "step": 150
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.872531418312388e-06,
+      "loss": 1.2103,
+      "step": 160
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.0520646319569123e-06,
+      "loss": 1.187,
+      "step": 170
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.2315978456014364e-06,
+      "loss": 1.1591,
+      "step": 180
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 3.411131059245961e-06,
+      "loss": 1.144,
+      "step": 190
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 3.590664272890485e-06,
+      "loss": 1.0904,
+      "step": 200
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 3.770197486535009e-06,
+      "loss": 1.0401,
+      "step": 210
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 3.949730700179533e-06,
+      "loss": 1.0281,
+      "step": 220
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.129263913824058e-06,
+      "loss": 0.9577,
+      "step": 230
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.3087971274685824e-06,
+      "loss": 0.9251,
+      "step": 240
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.488330341113106e-06,
+      "loss": 0.8964,
+      "step": 250
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.66786355475763e-06,
+      "loss": 0.8897,
+      "step": 260
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.847396768402154e-06,
+      "loss": 0.8481,
+      "step": 270
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.026929982046679e-06,
+      "loss": 0.8048,
+      "step": 280
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.206463195691203e-06,
+      "loss": 0.8014,
+      "step": 290
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.385996409335727e-06,
+      "loss": 0.7511,
+      "step": 300
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.565529622980251e-06,
+      "loss": 0.741,
+      "step": 310
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.745062836624776e-06,
+      "loss": 0.7327,
+      "step": 320
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.9245960502693004e-06,
+      "loss": 0.7492,
+      "step": 330
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.1041292639138246e-06,
+      "loss": 0.7613,
+      "step": 340
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.283662477558349e-06,
+      "loss": 0.6902,
+      "step": 350
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.463195691202873e-06,
+      "loss": 0.6687,
+      "step": 360
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.642728904847396e-06,
+      "loss": 0.6623,
+      "step": 370
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.822262118491922e-06,
+      "loss": 0.6901,
+      "step": 380
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 7.001795332136446e-06,
+      "loss": 0.6638,
+      "step": 390
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 7.18132854578097e-06,
+      "loss": 0.6414,
+      "step": 400
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 7.360861759425494e-06,
+      "loss": 0.6877,
+      "step": 410
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 7.540394973070018e-06,
+      "loss": 0.6249,
+      "step": 420
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 7.719928186714543e-06,
+      "loss": 0.6099,
+      "step": 430
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.899461400359067e-06,
+      "loss": 0.6321,
+      "step": 440
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 8.07899461400359e-06,
+      "loss": 0.6431,
+      "step": 450
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.258527827648117e-06,
+      "loss": 0.6404,
+      "step": 460
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.43806104129264e-06,
+      "loss": 0.6148,
+      "step": 470
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 8.617594254937165e-06,
+      "loss": 0.5844,
+      "step": 480
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.797127468581689e-06,
+      "loss": 0.6316,
+      "step": 490
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.976660682226211e-06,
+      "loss": 0.6018,
+      "step": 500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 9.156193895870736e-06,
+      "loss": 0.5843,
+      "step": 510
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.33572710951526e-06,
+      "loss": 0.5986,
+      "step": 520
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 9.515260323159784e-06,
+      "loss": 0.5718,
+      "step": 530
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.694793536804308e-06,
+      "loss": 0.5163,
+      "step": 540
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.874326750448834e-06,
+      "loss": 0.5448,
+      "step": 550
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.8387530836510428
+      },
+      "eval_f1": {
+        "f1": 0.8242528613153339
+      },
+      "eval_loss": 0.43474748730659485,
+      "eval_precision": {
+        "precision": 0.8238035334510331
+      },
+      "eval_recall": {
+        "recall": 0.8266051811271123
+      },
+      "eval_runtime": 171.551,
+      "eval_samples_per_second": 103.969,
+      "eval_steps_per_second": 3.253,
+      "step": 557
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0053859964093358e-05,
+      "loss": 0.58,
+      "step": 560
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0233393177737882e-05,
+      "loss": 0.5298,
+      "step": 570
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.0412926391382406e-05,
+      "loss": 0.5592,
+      "step": 580
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.059245960502693e-05,
+      "loss": 0.5352,
+      "step": 590
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.0771992818671454e-05,
+      "loss": 0.5683,
+      "step": 600
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.0951526032315979e-05,
+      "loss": 0.571,
+      "step": 610
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.1131059245960503e-05,
+      "loss": 0.5603,
+      "step": 620
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.1310592459605028e-05,
+      "loss": 0.5157,
+      "step": 630
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.1490125673249553e-05,
+      "loss": 0.4885,
+      "step": 640
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.1669658886894077e-05,
+      "loss": 0.5005,
+      "step": 650
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.1849192100538601e-05,
+      "loss": 0.4951,
+      "step": 660
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.2028725314183125e-05,
+      "loss": 0.5197,
+      "step": 670
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.2208258527827649e-05,
+      "loss": 0.4746,
+      "step": 680
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.2387791741472173e-05,
+      "loss": 0.486,
+      "step": 690
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.2567324955116697e-05,
+      "loss": 0.4932,
+      "step": 700
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.2746858168761221e-05,
+      "loss": 0.5222,
+      "step": 710
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.2926391382405746e-05,
+      "loss": 0.434,
+      "step": 720
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.310592459605027e-05,
+      "loss": 0.4817,
+      "step": 730
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.3285457809694792e-05,
+      "loss": 0.487,
+      "step": 740
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.3464991023339318e-05,
+      "loss": 0.4478,
+      "step": 750
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.3644524236983844e-05,
+      "loss": 0.4564,
+      "step": 760
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.3824057450628366e-05,
+      "loss": 0.4455,
+      "step": 770
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.4003590664272892e-05,
+      "loss": 0.4701,
+      "step": 780
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.4183123877917415e-05,
+      "loss": 0.4333,
+      "step": 790
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.436265709156194e-05,
+      "loss": 0.4409,
+      "step": 800
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.4542190305206463e-05,
+      "loss": 0.4639,
+      "step": 810
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.4721723518850989e-05,
+      "loss": 0.4552,
+      "step": 820
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.4901256732495511e-05,
+      "loss": 0.4759,
+      "step": 830
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.5080789946140037e-05,
+      "loss": 0.4584,
+      "step": 840
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.5260323159784563e-05,
+      "loss": 0.4502,
+      "step": 850
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.5439856373429085e-05,
+      "loss": 0.4519,
+      "step": 860
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.561938958707361e-05,
+      "loss": 0.4695,
+      "step": 870
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.5798922800718133e-05,
+      "loss": 0.4466,
+      "step": 880
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.597845601436266e-05,
+      "loss": 0.4531,
+      "step": 890
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.615798922800718e-05,
+      "loss": 0.4674,
+      "step": 900
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.6337522441651707e-05,
+      "loss": 0.4262,
+      "step": 910
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.6517055655296233e-05,
+      "loss": 0.3945,
+      "step": 920
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6696588868940756e-05,
+      "loss": 0.4318,
+      "step": 930
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.687612208258528e-05,
+      "loss": 0.439,
+      "step": 940
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.7055655296229804e-05,
+      "loss": 0.4036,
+      "step": 950
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.723518850987433e-05,
+      "loss": 0.422,
+      "step": 960
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.7414721723518852e-05,
+      "loss": 0.4315,
+      "step": 970
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.7594254937163378e-05,
+      "loss": 0.4516,
+      "step": 980
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.77737881508079e-05,
+      "loss": 0.4092,
+      "step": 990
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.7953321364452423e-05,
+      "loss": 0.4367,
+      "step": 1000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.813285457809695e-05,
+      "loss": 0.4321,
+      "step": 1010
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.831238779174147e-05,
+      "loss": 0.3859,
+      "step": 1020
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.8491921005385997e-05,
+      "loss": 0.415,
+      "step": 1030
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.867145421903052e-05,
+      "loss": 0.4388,
+      "step": 1040
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8850987432675045e-05,
+      "loss": 0.3968,
+      "step": 1050
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.9030520646319568e-05,
+      "loss": 0.4172,
+      "step": 1060
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.9210053859964093e-05,
+      "loss": 0.4593,
+      "step": 1070
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.9389587073608616e-05,
+      "loss": 0.3924,
+      "step": 1080
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.9569120287253142e-05,
+      "loss": 0.4067,
+      "step": 1090
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.9748653500897668e-05,
+      "loss": 0.3906,
+      "step": 1100
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.992818671454219e-05,
+      "loss": 0.3915,
+      "step": 1110
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8952679973088137
+      },
+      "eval_f1": {
+        "f1": 0.8895212539619177
+      },
+      "eval_loss": 0.2781514525413513,
+      "eval_precision": {
+        "precision": 0.8895590270096669
+      },
+      "eval_recall": {
+        "recall": 0.892156947982043
+      },
+      "eval_runtime": 99.0662,
+      "eval_samples_per_second": 180.041,
+      "eval_steps_per_second": 5.633,
+      "step": 1115
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.0107719928186716e-05,
+      "loss": 0.4099,
+      "step": 1120
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.0287253141831238e-05,
+      "loss": 0.4165,
+      "step": 1130
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.0466786355475764e-05,
+      "loss": 0.3958,
+      "step": 1140
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.0646319569120286e-05,
+      "loss": 0.4353,
+      "step": 1150
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.0825852782764812e-05,
+      "loss": 0.3832,
+      "step": 1160
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.1005385996409338e-05,
+      "loss": 0.3967,
+      "step": 1170
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.118491921005386e-05,
+      "loss": 0.3614,
+      "step": 1180
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.1364452423698386e-05,
+      "loss": 0.3665,
+      "step": 1190
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.154398563734291e-05,
+      "loss": 0.3871,
+      "step": 1200
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.1723518850987435e-05,
+      "loss": 0.3748,
+      "step": 1210
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.1903052064631957e-05,
+      "loss": 0.396,
+      "step": 1220
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.2082585278276483e-05,
+      "loss": 0.374,
+      "step": 1230
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.2262118491921005e-05,
+      "loss": 0.3735,
+      "step": 1240
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.244165170556553e-05,
+      "loss": 0.3894,
+      "step": 1250
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.2621184919210057e-05,
+      "loss": 0.3443,
+      "step": 1260
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.280071813285458e-05,
+      "loss": 0.3402,
+      "step": 1270
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.2980251346499105e-05,
+      "loss": 0.4164,
+      "step": 1280
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.3159784560143628e-05,
+      "loss": 0.3879,
+      "step": 1290
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.3339317773788153e-05,
+      "loss": 0.3719,
+      "step": 1300
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.3518850987432676e-05,
+      "loss": 0.3477,
+      "step": 1310
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.3698384201077202e-05,
+      "loss": 0.3744,
+      "step": 1320
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.3877917414721724e-05,
+      "loss": 0.3833,
+      "step": 1330
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.405745062836625e-05,
+      "loss": 0.398,
+      "step": 1340
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.4236983842010776e-05,
+      "loss": 0.3562,
+      "step": 1350
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.4416517055655298e-05,
+      "loss": 0.3614,
+      "step": 1360
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.4596050269299824e-05,
+      "loss": 0.315,
+      "step": 1370
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.4775583482944346e-05,
+      "loss": 0.3441,
+      "step": 1380
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.495511669658887e-05,
+      "loss": 0.3589,
+      "step": 1390
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.5134649910233395e-05,
+      "loss": 0.3418,
+      "step": 1400
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.5314183123877917e-05,
+      "loss": 0.3696,
+      "step": 1410
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.5493716337522443e-05,
+      "loss": 0.3606,
+      "step": 1420
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.5673249551166965e-05,
+      "loss": 0.3389,
+      "step": 1430
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.585278276481149e-05,
+      "loss": 0.3375,
+      "step": 1440
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.6032315978456017e-05,
+      "loss": 0.3193,
+      "step": 1450
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.621184919210054e-05,
+      "loss": 0.3683,
+      "step": 1460
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.6391382405745062e-05,
+      "loss": 0.3203,
+      "step": 1470
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.6570915619389584e-05,
+      "loss": 0.3792,
+      "step": 1480
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.6750448833034114e-05,
+      "loss": 0.3343,
+      "step": 1490
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.6929982046678636e-05,
+      "loss": 0.3671,
+      "step": 1500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.710951526032316e-05,
+      "loss": 0.4002,
+      "step": 1510
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.7289048473967688e-05,
+      "loss": 0.3246,
+      "step": 1520
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.746858168761221e-05,
+      "loss": 0.3731,
+      "step": 1530
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.7648114901256732e-05,
+      "loss": 0.3652,
+      "step": 1540
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.7827648114901255e-05,
+      "loss": 0.3767,
+      "step": 1550
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.8007181328545784e-05,
+      "loss": 0.3135,
+      "step": 1560
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.8186714542190307e-05,
+      "loss": 0.3303,
+      "step": 1570
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.836624775583483e-05,
+      "loss": 0.3485,
+      "step": 1580
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.8545780969479358e-05,
+      "loss": 0.3528,
+      "step": 1590
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.872531418312388e-05,
+      "loss": 0.3205,
+      "step": 1600
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.8904847396768403e-05,
+      "loss": 0.3047,
+      "step": 1610
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.9084380610412926e-05,
+      "loss": 0.3656,
+      "step": 1620
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.9263913824057455e-05,
+      "loss": 0.3317,
+      "step": 1630
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.9443447037701977e-05,
+      "loss": 0.3318,
+      "step": 1640
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.96229802513465e-05,
+      "loss": 0.365,
+      "step": 1650
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.9802513464991022e-05,
+      "loss": 0.3273,
+      "step": 1660
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.998204667863555e-05,
+      "loss": 0.3246,
+      "step": 1670
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.9295245570755775
+      },
+      "eval_f1": {
+        "f1": 0.9243811536773696
+      },
+      "eval_loss": 0.19212597608566284,
+      "eval_precision": {
+        "precision": 0.9273415557649412
+      },
+      "eval_recall": {
+        "recall": 0.9219909309089409
+      },
+      "eval_runtime": 99.0388,
+      "eval_samples_per_second": 180.091,
+      "eval_steps_per_second": 5.634,
+      "step": 1672
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.0161579892280074e-05,
+      "loss": 0.3135,
+      "step": 1680
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.0341113105924596e-05,
+      "loss": 0.3145,
+      "step": 1690
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.0520646319569125e-05,
+      "loss": 0.3395,
+      "step": 1700
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.070017953321365e-05,
+      "loss": 0.2927,
+      "step": 1710
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.087971274685817e-05,
+      "loss": 0.3363,
+      "step": 1720
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.105924596050269e-05,
+      "loss": 0.3392,
+      "step": 1730
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.123877917414722e-05,
+      "loss": 0.3279,
+      "step": 1740
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.1418312387791744e-05,
+      "loss": 0.3194,
+      "step": 1750
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.159784560143627e-05,
+      "loss": 0.3206,
+      "step": 1760
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.1777378815080796e-05,
+      "loss": 0.3246,
+      "step": 1770
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.195691202872532e-05,
+      "loss": 0.3155,
+      "step": 1780
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.213644524236984e-05,
+      "loss": 0.301,
+      "step": 1790
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.231597845601436e-05,
+      "loss": 0.3287,
+      "step": 1800
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.249551166965889e-05,
+      "loss": 0.3282,
+      "step": 1810
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.2675044883303415e-05,
+      "loss": 0.2872,
+      "step": 1820
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.285457809694794e-05,
+      "loss": 0.3003,
+      "step": 1830
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.3034111310592466e-05,
+      "loss": 0.3009,
+      "step": 1840
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.321364452423699e-05,
+      "loss": 0.325,
+      "step": 1850
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.339317773788151e-05,
+      "loss": 0.3308,
+      "step": 1860
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.3572710951526034e-05,
+      "loss": 0.3175,
+      "step": 1870
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.375224416517056e-05,
+      "loss": 0.291,
+      "step": 1880
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.3931777378815085e-05,
+      "loss": 0.2936,
+      "step": 1890
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.411131059245961e-05,
+      "loss": 0.3304,
+      "step": 1900
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.429084380610413e-05,
+      "loss": 0.331,
+      "step": 1910
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.447037701974866e-05,
+      "loss": 0.3158,
+      "step": 1920
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.464991023339318e-05,
+      "loss": 0.3054,
+      "step": 1930
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.4829443447037704e-05,
+      "loss": 0.3063,
+      "step": 1940
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.500897666068223e-05,
+      "loss": 0.279,
+      "step": 1950
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.5188509874326756e-05,
+      "loss": 0.2986,
+      "step": 1960
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.536804308797128e-05,
+      "loss": 0.3184,
+      "step": 1970
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.55475763016158e-05,
+      "loss": 0.3623,
+      "step": 1980
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.572710951526032e-05,
+      "loss": 0.3178,
+      "step": 1990
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.5906642728904846e-05,
+      "loss": 0.2553,
+      "step": 2000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.6086175942549375e-05,
+      "loss": 0.3116,
+      "step": 2010
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.62657091561939e-05,
+      "loss": 0.3278,
+      "step": 2020
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.644524236983842e-05,
+      "loss": 0.311,
+      "step": 2030
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.662477558348294e-05,
+      "loss": 0.3116,
+      "step": 2040
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.6804308797127465e-05,
+      "loss": 0.3087,
+      "step": 2050
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.6983842010771994e-05,
+      "loss": 0.3091,
+      "step": 2060
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.7163375224416516e-05,
+      "loss": 0.2855,
+      "step": 2070
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.734290843806104e-05,
+      "loss": 0.313,
+      "step": 2080
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.752244165170557e-05,
+      "loss": 0.2992,
+      "step": 2090
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.770197486535009e-05,
+      "loss": 0.2946,
+      "step": 2100
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.788150807899461e-05,
+      "loss": 0.3034,
+      "step": 2110
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.8061041292639135e-05,
+      "loss": 0.2868,
+      "step": 2120
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.8240574506283664e-05,
+      "loss": 0.2818,
+      "step": 2130
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.842010771992819e-05,
+      "loss": 0.2943,
+      "step": 2140
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.859964093357271e-05,
+      "loss": 0.3166,
+      "step": 2150
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.877917414721723e-05,
+      "loss": 0.2947,
+      "step": 2160
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.895870736086176e-05,
+      "loss": 0.2764,
+      "step": 2170
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.9138240574506283e-05,
+      "loss": 0.278,
+      "step": 2180
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.9317773788150806e-05,
+      "loss": 0.3091,
+      "step": 2190
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.9497307001795335e-05,
+      "loss": 0.3336,
+      "step": 2200
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.967684021543986e-05,
+      "loss": 0.3256,
+      "step": 2210
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.985637342908438e-05,
+      "loss": 0.2664,
+      "step": 2220
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.00359066427289e-05,
+      "loss": 0.2819,
+      "step": 2230
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.9389437093518727
+      },
+      "eval_f1": {
+        "f1": 0.934636026587255
+      },
+      "eval_loss": 0.16309776902198792,
+      "eval_precision": {
+        "precision": 0.9384931858190604
+      },
+      "eval_recall": {
+        "recall": 0.9317771824858079
+      },
+      "eval_runtime": 98.5849,
+      "eval_samples_per_second": 180.92,
+      "eval_steps_per_second": 5.66,
+      "step": 2230
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.021543985637343e-05,
+      "loss": 0.2725,
+      "step": 2240
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.0394973070017954e-05,
+      "loss": 0.3239,
+      "step": 2250
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.0574506283662476e-05,
+      "loss": 0.2761,
+      "step": 2260
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.0754039497307006e-05,
+      "loss": 0.275,
+      "step": 2270
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.093357271095153e-05,
+      "loss": 0.2789,
+      "step": 2280
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.111310592459605e-05,
+      "loss": 0.2702,
+      "step": 2290
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.129263913824057e-05,
+      "loss": 0.3244,
+      "step": 2300
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.14721723518851e-05,
+      "loss": 0.3254,
+      "step": 2310
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.1651705565529625e-05,
+      "loss": 0.2969,
+      "step": 2320
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.183123877917415e-05,
+      "loss": 0.2545,
+      "step": 2330
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.2010771992818676e-05,
+      "loss": 0.2307,
+      "step": 2340
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.21903052064632e-05,
+      "loss": 0.3132,
+      "step": 2350
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.236983842010772e-05,
+      "loss": 0.3017,
+      "step": 2360
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.2549371633752244e-05,
+      "loss": 0.2885,
+      "step": 2370
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.272890484739677e-05,
+      "loss": 0.3085,
+      "step": 2380
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.2908438061041295e-05,
+      "loss": 0.3079,
+      "step": 2390
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.308797127468582e-05,
+      "loss": 0.2749,
+      "step": 2400
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.326750448833034e-05,
+      "loss": 0.2863,
+      "step": 2410
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.344703770197487e-05,
+      "loss": 0.2523,
+      "step": 2420
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.362657091561939e-05,
+      "loss": 0.2441,
+      "step": 2430
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.3806104129263914e-05,
+      "loss": 0.3174,
+      "step": 2440
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.398563734290844e-05,
+      "loss": 0.2938,
+      "step": 2450
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.4165170556552966e-05,
+      "loss": 0.2632,
+      "step": 2460
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.434470377019749e-05,
+      "loss": 0.2924,
+      "step": 2470
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.452423698384201e-05,
+      "loss": 0.3018,
+      "step": 2480
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.470377019748654e-05,
+      "loss": 0.2488,
+      "step": 2490
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.488330341113106e-05,
+      "loss": 0.2659,
+      "step": 2500
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.5062836624775585e-05,
+      "loss": 0.2689,
+      "step": 2510
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.5242369838420114e-05,
+      "loss": 0.2702,
+      "step": 2520
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.5421903052064636e-05,
+      "loss": 0.2695,
+      "step": 2530
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.560143626570916e-05,
+      "loss": 0.2629,
+      "step": 2540
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.578096947935368e-05,
+      "loss": 0.2931,
+      "step": 2550
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.596050269299821e-05,
+      "loss": 0.2865,
+      "step": 2560
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.614003590664273e-05,
+      "loss": 0.2751,
+      "step": 2570
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6319569120287255e-05,
+      "loss": 0.2565,
+      "step": 2580
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.6499102333931784e-05,
+      "loss": 0.2589,
+      "step": 2590
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.667863554757631e-05,
+      "loss": 0.2501,
+      "step": 2600
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.685816876122083e-05,
+      "loss": 0.2524,
+      "step": 2610
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.703770197486535e-05,
+      "loss": 0.2596,
+      "step": 2620
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.721723518850988e-05,
+      "loss": 0.2567,
+      "step": 2630
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.7396768402154403e-05,
+      "loss": 0.2886,
+      "step": 2640
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.7576301615798926e-05,
+      "loss": 0.2879,
+      "step": 2650
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.775583482944345e-05,
+      "loss": 0.2703,
+      "step": 2660
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.793536804308798e-05,
+      "loss": 0.3013,
+      "step": 2670
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.81149012567325e-05,
+      "loss": 0.2801,
+      "step": 2680
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.829443447037702e-05,
+      "loss": 0.2886,
+      "step": 2690
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.847396768402155e-05,
+      "loss": 0.3084,
+      "step": 2700
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.8653500897666074e-05,
+      "loss": 0.2886,
+      "step": 2710
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.8833034111310596e-05,
+      "loss": 0.2678,
+      "step": 2720
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.901256732495512e-05,
+      "loss": 0.2628,
+      "step": 2730
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.919210053859965e-05,
+      "loss": 0.2834,
+      "step": 2740
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.937163375224417e-05,
+      "loss": 0.2775,
+      "step": 2750
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.955116696588869e-05,
+      "loss": 0.2987,
+      "step": 2760
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.9730700179533215e-05,
+      "loss": 0.2409,
+      "step": 2770
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.991023339317774e-05,
+      "loss": 0.3022,
+      "step": 2780
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.9471294012110338
+      },
+      "eval_f1": {
+        "f1": 0.944197644580418
+      },
+      "eval_loss": 0.1480465531349182,
+      "eval_precision": {
+        "precision": 0.9443621962369058
+      },
+      "eval_recall": {
+        "recall": 0.9440973623586151
+      },
+      "eval_runtime": 99.3848,
+      "eval_samples_per_second": 179.464,
+      "eval_steps_per_second": 5.615,
+      "step": 2787
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 4.999002593257531e-05,
+      "loss": 0.2842,
+      "step": 2790
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.9970077797725914e-05,
+      "loss": 0.2321,
+      "step": 2800
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.995012966287652e-05,
+      "loss": 0.259,
+      "step": 2810
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.993018152802713e-05,
+      "loss": 0.2541,
+      "step": 2820
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.991023339317774e-05,
+      "loss": 0.2419,
+      "step": 2830
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.989028525832835e-05,
+      "loss": 0.2506,
+      "step": 2840
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.9870337123478956e-05,
+      "loss": 0.24,
+      "step": 2850
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.985038898862957e-05,
+      "loss": 0.2443,
+      "step": 2860
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.9830440853780174e-05,
+      "loss": 0.2645,
+      "step": 2870
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.9810492718930786e-05,
+      "loss": 0.2679,
+      "step": 2880
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.979054458408139e-05,
+      "loss": 0.2738,
+      "step": 2890
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.9770596449232e-05,
+      "loss": 0.2329,
+      "step": 2900
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.975064831438261e-05,
+      "loss": 0.2766,
+      "step": 2910
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.9730700179533215e-05,
+      "loss": 0.243,
+      "step": 2920
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.971075204468382e-05,
+      "loss": 0.2722,
+      "step": 2930
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.969080390983443e-05,
+      "loss": 0.2556,
+      "step": 2940
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.967085577498504e-05,
+      "loss": 0.2544,
+      "step": 2950
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.965090764013565e-05,
+      "loss": 0.2353,
+      "step": 2960
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.963095950528626e-05,
+      "loss": 0.252,
+      "step": 2970
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.961101137043687e-05,
+      "loss": 0.2303,
+      "step": 2980
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.9591063235587475e-05,
+      "loss": 0.2357,
+      "step": 2990
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.957111510073809e-05,
+      "loss": 0.2529,
+      "step": 3000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.955116696588869e-05,
+      "loss": 0.2381,
+      "step": 3010
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.95312188310393e-05,
+      "loss": 0.2616,
+      "step": 3020
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.951127069618991e-05,
+      "loss": 0.2836,
+      "step": 3030
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.9491322561340517e-05,
+      "loss": 0.2726,
+      "step": 3040
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.947137442649112e-05,
+      "loss": 0.2487,
+      "step": 3050
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.9451426291641735e-05,
+      "loss": 0.2577,
+      "step": 3060
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.943147815679234e-05,
+      "loss": 0.2453,
+      "step": 3070
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.941153002194295e-05,
+      "loss": 0.2682,
+      "step": 3080
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.939158188709356e-05,
+      "loss": 0.2463,
+      "step": 3090
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.937163375224417e-05,
+      "loss": 0.254,
+      "step": 3100
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.9351685617394776e-05,
+      "loss": 0.2516,
+      "step": 3110
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.933173748254539e-05,
+      "loss": 0.2395,
+      "step": 3120
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.9311789347695994e-05,
+      "loss": 0.2375,
+      "step": 3130
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.92918412128466e-05,
+      "loss": 0.2551,
+      "step": 3140
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.927189307799721e-05,
+      "loss": 0.2391,
+      "step": 3150
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.925194494314782e-05,
+      "loss": 0.2501,
+      "step": 3160
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.923199680829842e-05,
+      "loss": 0.248,
+      "step": 3170
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.9212048673449036e-05,
+      "loss": 0.2357,
+      "step": 3180
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.919210053859965e-05,
+      "loss": 0.2454,
+      "step": 3190
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.917215240375025e-05,
+      "loss": 0.2728,
+      "step": 3200
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.915220426890086e-05,
+      "loss": 0.2789,
+      "step": 3210
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.913225613405147e-05,
+      "loss": 0.2559,
+      "step": 3220
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.911230799920208e-05,
+      "loss": 0.2631,
+      "step": 3230
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.909235986435268e-05,
+      "loss": 0.2292,
+      "step": 3240
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.9072411729503295e-05,
+      "loss": 0.2301,
+      "step": 3250
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.90524635946539e-05,
+      "loss": 0.2711,
+      "step": 3260
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.903251545980451e-05,
+      "loss": 0.2588,
+      "step": 3270
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.901256732495512e-05,
+      "loss": 0.2683,
+      "step": 3280
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.8992619190105724e-05,
+      "loss": 0.1896,
+      "step": 3290
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.897267105525634e-05,
+      "loss": 0.2639,
+      "step": 3300
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.895272292040695e-05,
+      "loss": 0.2606,
+      "step": 3310
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.893277478555755e-05,
+      "loss": 0.2751,
+      "step": 3320
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.891282665070816e-05,
+      "loss": 0.2423,
+      "step": 3330
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.889287851585877e-05,
+      "loss": 0.2283,
+      "step": 3340
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.9510540479928234
+      },
+      "eval_f1": {
+        "f1": 0.9482048430595941
+      },
+      "eval_loss": 0.13114869594573975,
+      "eval_precision": {
+        "precision": 0.9513484013297254
+      },
+      "eval_recall": {
+        "recall": 0.9456754807307988
+      },
+      "eval_runtime": 99.7361,
+      "eval_samples_per_second": 178.832,
+      "eval_steps_per_second": 5.595,
+      "step": 3345
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.887293038100938e-05,
+      "loss": 0.2306,
+      "step": 3350
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.8852982246159984e-05,
+      "loss": 0.2259,
+      "step": 3360
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.8833034111310596e-05,
+      "loss": 0.2353,
+      "step": 3370
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.88130859764612e-05,
+      "loss": 0.2264,
+      "step": 3380
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.8793137841611814e-05,
+      "loss": 0.2264,
+      "step": 3390
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.877318970676242e-05,
+      "loss": 0.2403,
+      "step": 3400
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.8753241571913026e-05,
+      "loss": 0.2476,
+      "step": 3410
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.873329343706364e-05,
+      "loss": 0.2415,
+      "step": 3420
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.871334530221425e-05,
+      "loss": 0.26,
+      "step": 3430
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.869339716736485e-05,
+      "loss": 0.2659,
+      "step": 3440
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.867344903251546e-05,
+      "loss": 0.2114,
+      "step": 3450
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.8653500897666074e-05,
+      "loss": 0.2485,
+      "step": 3460
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.863355276281668e-05,
+      "loss": 0.2247,
+      "step": 3470
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.8613604627967285e-05,
+      "loss": 0.2595,
+      "step": 3480
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.85936564931179e-05,
+      "loss": 0.2193,
+      "step": 3490
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.85737083582685e-05,
+      "loss": 0.2051,
+      "step": 3500
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.8553760223419116e-05,
+      "loss": 0.2105,
+      "step": 3510
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.853381208856972e-05,
+      "loss": 0.2104,
+      "step": 3520
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.851386395372033e-05,
+      "loss": 0.2595,
+      "step": 3530
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.849391581887094e-05,
+      "loss": 0.2199,
+      "step": 3540
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.847396768402155e-05,
+      "loss": 0.2463,
+      "step": 3550
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.845401954917215e-05,
+      "loss": 0.221,
+      "step": 3560
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.843407141432276e-05,
+      "loss": 0.2207,
+      "step": 3570
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.8414123279473375e-05,
+      "loss": 0.2319,
+      "step": 3580
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.839417514462398e-05,
+      "loss": 0.2333,
+      "step": 3590
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.8374227009774586e-05,
+      "loss": 0.237,
+      "step": 3600
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.83542788749252e-05,
+      "loss": 0.2096,
+      "step": 3610
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.8334330740075804e-05,
+      "loss": 0.236,
+      "step": 3620
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.831438260522642e-05,
+      "loss": 0.2376,
+      "step": 3630
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.829443447037702e-05,
+      "loss": 0.251,
+      "step": 3640
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.827448633552763e-05,
+      "loss": 0.1974,
+      "step": 3650
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.825453820067824e-05,
+      "loss": 0.2334,
+      "step": 3660
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.823459006582885e-05,
+      "loss": 0.2289,
+      "step": 3670
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.821464193097945e-05,
+      "loss": 0.2252,
+      "step": 3680
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.8194693796130064e-05,
+      "loss": 0.239,
+      "step": 3690
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.8174745661280676e-05,
+      "loss": 0.2166,
+      "step": 3700
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.815479752643128e-05,
+      "loss": 0.225,
+      "step": 3710
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.813484939158189e-05,
+      "loss": 0.2038,
+      "step": 3720
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.81149012567325e-05,
+      "loss": 0.214,
+      "step": 3730
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.8094953121883106e-05,
+      "loss": 0.1958,
+      "step": 3740
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.807500498703372e-05,
+      "loss": 0.2275,
+      "step": 3750
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.8055056852184324e-05,
+      "loss": 0.2324,
+      "step": 3760
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.803510871733493e-05,
+      "loss": 0.2202,
+      "step": 3770
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.801516058248554e-05,
+      "loss": 0.2285,
+      "step": 3780
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.7995212447636154e-05,
+      "loss": 0.2385,
+      "step": 3790
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.797526431278675e-05,
+      "loss": 0.2366,
+      "step": 3800
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.7955316177937365e-05,
+      "loss": 0.2221,
+      "step": 3810
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.793536804308798e-05,
+      "loss": 0.2349,
+      "step": 3820
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.791541990823858e-05,
+      "loss": 0.2749,
+      "step": 3830
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.789547177338919e-05,
+      "loss": 0.2168,
+      "step": 3840
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.78755236385398e-05,
+      "loss": 0.1872,
+      "step": 3850
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.785557550369041e-05,
+      "loss": 0.2502,
+      "step": 3860
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.783562736884102e-05,
+      "loss": 0.1964,
+      "step": 3870
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.7815679233991625e-05,
+      "loss": 0.2425,
+      "step": 3880
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.779573109914223e-05,
+      "loss": 0.2481,
+      "step": 3890
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.777578296429284e-05,
+      "loss": 0.2376,
+      "step": 3900
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.9595761381475667
+      },
+      "eval_f1": {
+        "f1": 0.95694652190804
+      },
+      "eval_loss": 0.10949217528104782,
+      "eval_precision": {
+        "precision": 0.9583449741949119
+      },
+      "eval_recall": {
+        "recall": 0.9557128312494094
+      },
+      "eval_runtime": 99.0631,
+      "eval_samples_per_second": 180.047,
+      "eval_steps_per_second": 5.633,
+      "step": 3902
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.775583482944345e-05,
+      "loss": 0.2048,
+      "step": 3910
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.7735886694594054e-05,
+      "loss": 0.2573,
+      "step": 3920
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.7715938559744666e-05,
+      "loss": 0.2261,
+      "step": 3930
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.769599042489528e-05,
+      "loss": 0.2094,
+      "step": 3940
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.767604229004588e-05,
+      "loss": 0.2148,
+      "step": 3950
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.765609415519649e-05,
+      "loss": 0.2023,
+      "step": 3960
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.76361460203471e-05,
+      "loss": 0.2031,
+      "step": 3970
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.761619788549771e-05,
+      "loss": 0.2051,
+      "step": 3980
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.7596249750648313e-05,
+      "loss": 0.1944,
+      "step": 3990
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.7576301615798926e-05,
+      "loss": 0.2319,
+      "step": 4000
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.755635348094953e-05,
+      "loss": 0.2164,
+      "step": 4010
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.7536405346100144e-05,
+      "loss": 0.1977,
+      "step": 4020
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.751645721125075e-05,
+      "loss": 0.2499,
+      "step": 4030
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.7496509076401355e-05,
+      "loss": 0.1717,
+      "step": 4040
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.747656094155197e-05,
+      "loss": 0.1886,
+      "step": 4050
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.745661280670258e-05,
+      "loss": 0.2132,
+      "step": 4060
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.743666467185318e-05,
+      "loss": 0.2238,
+      "step": 4070
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.741671653700379e-05,
+      "loss": 0.2213,
+      "step": 4080
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.7396768402154403e-05,
+      "loss": 0.1884,
+      "step": 4090
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.737682026730501e-05,
+      "loss": 0.2112,
+      "step": 4100
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.7356872132455615e-05,
+      "loss": 0.218,
+      "step": 4110
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.733692399760623e-05,
+      "loss": 0.1901,
+      "step": 4120
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.731697586275683e-05,
+      "loss": 0.2182,
+      "step": 4130
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.7297027727907445e-05,
+      "loss": 0.2061,
+      "step": 4140
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.727707959305805e-05,
+      "loss": 0.1767,
+      "step": 4150
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.7257131458208656e-05,
+      "loss": 0.2376,
+      "step": 4160
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.723718332335927e-05,
+      "loss": 0.203,
+      "step": 4170
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.721723518850988e-05,
+      "loss": 0.1879,
+      "step": 4180
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.719728705366048e-05,
+      "loss": 0.2355,
+      "step": 4190
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.717733891881109e-05,
+      "loss": 0.1916,
+      "step": 4200
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.7157390783961705e-05,
+      "loss": 0.2133,
+      "step": 4210
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.713744264911231e-05,
+      "loss": 0.208,
+      "step": 4220
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.7117494514262916e-05,
+      "loss": 0.2364,
+      "step": 4230
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.709754637941353e-05,
+      "loss": 0.2049,
+      "step": 4240
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.7077598244564134e-05,
+      "loss": 0.2296,
+      "step": 4250
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.7057650109714746e-05,
+      "loss": 0.221,
+      "step": 4260
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.703770197486535e-05,
+      "loss": 0.1917,
+      "step": 4270
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.701775384001596e-05,
+      "loss": 0.2183,
+      "step": 4280
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.699780570516657e-05,
+      "loss": 0.1869,
+      "step": 4290
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.697785757031718e-05,
+      "loss": 0.1975,
+      "step": 4300
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.695790943546778e-05,
+      "loss": 0.2274,
+      "step": 4310
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.693796130061839e-05,
+      "loss": 0.1973,
+      "step": 4320
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.6918013165769006e-05,
+      "loss": 0.2005,
+      "step": 4330
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.689806503091961e-05,
+      "loss": 0.1978,
+      "step": 4340
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.687811689607022e-05,
+      "loss": 0.1986,
+      "step": 4350
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.685816876122083e-05,
+      "loss": 0.2013,
+      "step": 4360
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.6838220626371435e-05,
+      "loss": 0.2218,
+      "step": 4370
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.681827249152205e-05,
+      "loss": 0.2168,
+      "step": 4380
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.679832435667265e-05,
+      "loss": 0.1709,
+      "step": 4390
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.677837622182326e-05,
+      "loss": 0.2249,
+      "step": 4400
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.675842808697387e-05,
+      "loss": 0.2246,
+      "step": 4410
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.673847995212448e-05,
+      "loss": 0.2163,
+      "step": 4420
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.671853181727508e-05,
+      "loss": 0.1922,
+      "step": 4430
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.6698583682425695e-05,
+      "loss": 0.2157,
+      "step": 4440
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.667863554757631e-05,
+      "loss": 0.2272,
+      "step": 4450
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.665868741272691e-05,
+      "loss": 0.2027,
+      "step": 4460
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.9548665620094191
+      },
+      "eval_f1": {
+        "f1": 0.951539693294557
+      },
+      "eval_loss": 0.12259286642074585,
+      "eval_precision": {
+        "precision": 0.9548733769718504
+      },
+      "eval_recall": {
+        "recall": 0.9494659029809656
+      },
+      "eval_runtime": 98.7615,
+      "eval_samples_per_second": 180.597,
+      "eval_steps_per_second": 5.65,
+      "step": 4460
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.663873927787752e-05,
+      "loss": 0.1895,
+      "step": 4470
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.661879114302813e-05,
+      "loss": 0.1628,
+      "step": 4480
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.6598843008178736e-05,
+      "loss": 0.1996,
+      "step": 4490
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.657889487332935e-05,
+      "loss": 0.2024,
+      "step": 4500
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.6558946738479954e-05,
+      "loss": 0.1887,
+      "step": 4510
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.653899860363056e-05,
+      "loss": 0.2224,
+      "step": 4520
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.651905046878117e-05,
+      "loss": 0.1992,
+      "step": 4530
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.6499102333931784e-05,
+      "loss": 0.2027,
+      "step": 4540
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.647915419908238e-05,
+      "loss": 0.1844,
+      "step": 4550
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.6459206064232996e-05,
+      "loss": 0.1838,
+      "step": 4560
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.643925792938361e-05,
+      "loss": 0.1897,
+      "step": 4570
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.6419309794534214e-05,
+      "loss": 0.1794,
+      "step": 4580
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.639936165968482e-05,
+      "loss": 0.2052,
+      "step": 4590
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.637941352483543e-05,
+      "loss": 0.225,
+      "step": 4600
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.635946538998604e-05,
+      "loss": 0.1905,
+      "step": 4610
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.633951725513665e-05,
+      "loss": 0.1933,
+      "step": 4620
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.6319569120287255e-05,
+      "loss": 0.2084,
+      "step": 4630
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.629962098543786e-05,
+      "loss": 0.1947,
+      "step": 4640
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.627967285058847e-05,
+      "loss": 0.1984,
+      "step": 4650
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.625972471573908e-05,
+      "loss": 0.21,
+      "step": 4660
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.6239776580889684e-05,
+      "loss": 0.2083,
+      "step": 4670
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.62198284460403e-05,
+      "loss": 0.2066,
+      "step": 4680
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.619988031119091e-05,
+      "loss": 0.1874,
+      "step": 4690
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.6179932176341515e-05,
+      "loss": 0.2062,
+      "step": 4700
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.615998404149212e-05,
+      "loss": 0.1714,
+      "step": 4710
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.614003590664273e-05,
+      "loss": 0.1864,
+      "step": 4720
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.612008777179334e-05,
+      "loss": 0.1997,
+      "step": 4730
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.6100139636943944e-05,
+      "loss": 0.2102,
+      "step": 4740
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.6080191502094556e-05,
+      "loss": 0.2008,
+      "step": 4750
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.606024336724516e-05,
+      "loss": 0.1874,
+      "step": 4760
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.6040295232395774e-05,
+      "loss": 0.1718,
+      "step": 4770
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.602034709754638e-05,
+      "loss": 0.1799,
+      "step": 4780
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.6000398962696986e-05,
+      "loss": 0.1741,
+      "step": 4790
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.59804508278476e-05,
+      "loss": 0.1786,
+      "step": 4800
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.596050269299821e-05,
+      "loss": 0.1853,
+      "step": 4810
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.5940554558148816e-05,
+      "loss": 0.1964,
+      "step": 4820
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.592060642329942e-05,
+      "loss": 0.2065,
+      "step": 4830
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.5900658288450034e-05,
+      "loss": 0.2021,
+      "step": 4840
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.588071015360064e-05,
+      "loss": 0.1887,
+      "step": 4850
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.5860762018751245e-05,
+      "loss": 0.2203,
+      "step": 4860
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.584081388390186e-05,
+      "loss": 0.1808,
+      "step": 4870
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.582086574905246e-05,
+      "loss": 0.2213,
+      "step": 4880
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.5800917614203076e-05,
+      "loss": 0.1934,
+      "step": 4890
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.578096947935368e-05,
+      "loss": 0.2048,
+      "step": 4900
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.576102134450429e-05,
+      "loss": 0.217,
+      "step": 4910
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.57410732096549e-05,
+      "loss": 0.2161,
+      "step": 4920
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.572112507480551e-05,
+      "loss": 0.2145,
+      "step": 4930
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.570117693995612e-05,
+      "loss": 0.1722,
+      "step": 4940
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.568122880510672e-05,
+      "loss": 0.1904,
+      "step": 4950
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.5661280670257335e-05,
+      "loss": 0.1888,
+      "step": 4960
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.564133253540794e-05,
+      "loss": 0.1931,
+      "step": 4970
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.5621384400558546e-05,
+      "loss": 0.1607,
+      "step": 4980
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.560143626570916e-05,
+      "loss": 0.2014,
+      "step": 4990
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.5581488130859764e-05,
+      "loss": 0.2043,
+      "step": 5000
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.556153999601038e-05,
+      "loss": 0.1807,
+      "step": 5010
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.9633886521641624
+      },
+      "eval_f1": {
+        "f1": 0.9611100614395742
+      },
+      "eval_loss": 0.10741779208183289,
+      "eval_precision": {
+        "precision": 0.9618083534790791
+      },
+      "eval_recall": {
+        "recall": 0.9605394028604899
+      },
+      "eval_runtime": 98.612,
+      "eval_samples_per_second": 180.87,
+      "eval_steps_per_second": 5.659,
+      "step": 5017
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.554159186116098e-05,
+      "loss": 0.1723,
+      "step": 5020
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.5521643726311595e-05,
+      "loss": 0.1929,
+      "step": 5030
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.55016955914622e-05,
+      "loss": 0.165,
+      "step": 5040
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.548174745661281e-05,
+      "loss": 0.202,
+      "step": 5050
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.546179932176342e-05,
+      "loss": 0.2107,
+      "step": 5060
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.5441851186914024e-05,
+      "loss": 0.1745,
+      "step": 5070
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.5421903052064636e-05,
+      "loss": 0.2109,
+      "step": 5080
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.540195491721524e-05,
+      "loss": 0.1916,
+      "step": 5090
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.538200678236585e-05,
+      "loss": 0.204,
+      "step": 5100
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.536205864751646e-05,
+      "loss": 0.1683,
+      "step": 5110
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.5342110512667066e-05,
+      "loss": 0.1848,
+      "step": 5120
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.532216237781768e-05,
+      "loss": 0.1638,
+      "step": 5130
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.5302214242968284e-05,
+      "loss": 0.1545,
+      "step": 5140
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.5282266108118896e-05,
+      "loss": 0.1886,
+      "step": 5150
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.52623179732695e-05,
+      "loss": 0.188,
+      "step": 5160
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.5242369838420114e-05,
+      "loss": 0.197,
+      "step": 5170
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.522242170357072e-05,
+      "loss": 0.1625,
+      "step": 5180
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.5202473568721325e-05,
+      "loss": 0.1942,
+      "step": 5190
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.518252543387194e-05,
+      "loss": 0.1781,
+      "step": 5200
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.516257729902254e-05,
+      "loss": 0.1661,
+      "step": 5210
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.514262916417315e-05,
+      "loss": 0.197,
+      "step": 5220
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.512268102932376e-05,
+      "loss": 0.193,
+      "step": 5230
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.510273289447437e-05,
+      "loss": 0.1884,
+      "step": 5240
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.508278475962498e-05,
+      "loss": 0.1829,
+      "step": 5250
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.5062836624775585e-05,
+      "loss": 0.1854,
+      "step": 5260
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.50428884899262e-05,
+      "loss": 0.2,
+      "step": 5270
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.50229403550768e-05,
+      "loss": 0.1769,
+      "step": 5280
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.5002992220227415e-05,
+      "loss": 0.1876,
+      "step": 5290
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.498304408537802e-05,
+      "loss": 0.1598,
+      "step": 5300
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.4963095950528626e-05,
+      "loss": 0.1857,
+      "step": 5310
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.494314781567924e-05,
+      "loss": 0.1982,
+      "step": 5320
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.4923199680829844e-05,
+      "loss": 0.1846,
+      "step": 5330
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.490325154598045e-05,
+      "loss": 0.1707,
+      "step": 5340
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.488330341113106e-05,
+      "loss": 0.1952,
+      "step": 5350
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.486335527628167e-05,
+      "loss": 0.1499,
+      "step": 5360
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.484340714143228e-05,
+      "loss": 0.1857,
+      "step": 5370
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.4823459006582886e-05,
+      "loss": 0.1557,
+      "step": 5380
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.48035108717335e-05,
+      "loss": 0.1716,
+      "step": 5390
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.4783562736884104e-05,
+      "loss": 0.1696,
+      "step": 5400
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.476361460203471e-05,
+      "loss": 0.1695,
+      "step": 5410
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.474366646718532e-05,
+      "loss": 0.2097,
+      "step": 5420
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.472371833233593e-05,
+      "loss": 0.1986,
+      "step": 5430
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.470377019748654e-05,
+      "loss": 0.1548,
+      "step": 5440
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.4683822062637145e-05,
+      "loss": 0.1907,
+      "step": 5450
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.466387392778775e-05,
+      "loss": 0.1706,
+      "step": 5460
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.4643925792938363e-05,
+      "loss": 0.1791,
+      "step": 5470
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.462397765808897e-05,
+      "loss": 0.1913,
+      "step": 5480
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.4604029523239575e-05,
+      "loss": 0.1779,
+      "step": 5490
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.458408138839019e-05,
+      "loss": 0.172,
+      "step": 5500
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.45641332535408e-05,
+      "loss": 0.1696,
+      "step": 5510
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.4544185118691405e-05,
+      "loss": 0.1666,
+      "step": 5520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.452423698384201e-05,
+      "loss": 0.1601,
+      "step": 5530
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.450428884899262e-05,
+      "loss": 0.1778,
+      "step": 5540
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.448434071414323e-05,
+      "loss": 0.1732,
+      "step": 5550
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.446439257929384e-05,
+      "loss": 0.1886,
+      "step": 5560
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.2238,
+      "step": 5570
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.967425431711146
+      },
+      "eval_f1": {
+        "f1": 0.9654836432017294
+      },
+      "eval_loss": 0.09295742213726044,
+      "eval_precision": {
+        "precision": 0.9661647785647793
+      },
+      "eval_recall": {
+        "recall": 0.9648241815791808
+      },
+      "eval_runtime": 98.8994,
+      "eval_samples_per_second": 180.345,
+      "eval_steps_per_second": 5.642,
+      "step": 5575
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 4.442449630959505e-05,
+      "loss": 0.1974,
+      "step": 5580
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 4.4404548174745665e-05,
+      "loss": 0.1731,
+      "step": 5590
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 4.438460003989628e-05,
+      "loss": 0.1664,
+      "step": 5600
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 4.4364651905046876e-05,
+      "loss": 0.1904,
+      "step": 5610
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 4.434470377019749e-05,
+      "loss": 0.1393,
+      "step": 5620
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 4.43247556353481e-05,
+      "loss": 0.1697,
+      "step": 5630
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 4.4304807500498706e-05,
+      "loss": 0.1689,
+      "step": 5640
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.428485936564931e-05,
+      "loss": 0.1643,
+      "step": 5650
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 4.4264911230799924e-05,
+      "loss": 0.1638,
+      "step": 5660
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 4.424496309595053e-05,
+      "loss": 0.1539,
+      "step": 5670
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 4.422501496110114e-05,
+      "loss": 0.1588,
+      "step": 5680
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 4.420506682625175e-05,
+      "loss": 0.1461,
+      "step": 5690
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 4.418511869140235e-05,
+      "loss": 0.1629,
+      "step": 5700
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 4.4165170556552966e-05,
+      "loss": 0.1755,
+      "step": 5710
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 4.414522242170358e-05,
+      "loss": 0.1783,
+      "step": 5720
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 4.412527428685418e-05,
+      "loss": 0.1712,
+      "step": 5730
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 4.410532615200479e-05,
+      "loss": 0.1646,
+      "step": 5740
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 4.40853780171554e-05,
+      "loss": 0.1553,
+      "step": 5750
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 4.406542988230601e-05,
+      "loss": 0.1619,
+      "step": 5760
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 4.404548174745661e-05,
+      "loss": 0.1991,
+      "step": 5770
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 4.4025533612607225e-05,
+      "loss": 0.1806,
+      "step": 5780
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 4.400558547775783e-05,
+      "loss": 0.1616,
+      "step": 5790
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 4.398563734290844e-05,
+      "loss": 0.1912,
+      "step": 5800
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 4.396568920805905e-05,
+      "loss": 0.1764,
+      "step": 5810
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 4.3945741073209655e-05,
+      "loss": 0.1591,
+      "step": 5820
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 4.392579293836027e-05,
+      "loss": 0.1534,
+      "step": 5830
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 4.390584480351088e-05,
+      "loss": 0.1773,
+      "step": 5840
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 4.388589666866148e-05,
+      "loss": 0.1775,
+      "step": 5850
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 4.386594853381209e-05,
+      "loss": 0.1761,
+      "step": 5860
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 4.38460003989627e-05,
+      "loss": 0.1387,
+      "step": 5870
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 4.382605226411331e-05,
+      "loss": 0.2006,
+      "step": 5880
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 4.3806104129263914e-05,
+      "loss": 0.166,
+      "step": 5890
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 4.3786155994414526e-05,
+      "loss": 0.1741,
+      "step": 5900
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 4.376620785956513e-05,
+      "loss": 0.1627,
+      "step": 5910
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 4.3746259724715744e-05,
+      "loss": 0.1704,
+      "step": 5920
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 4.372631158986635e-05,
+      "loss": 0.1686,
+      "step": 5930
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 4.3706363455016956e-05,
+      "loss": 0.1717,
+      "step": 5940
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 4.368641532016757e-05,
+      "loss": 0.1795,
+      "step": 5950
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 4.366646718531818e-05,
+      "loss": 0.196,
+      "step": 5960
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 4.364651905046878e-05,
+      "loss": 0.1749,
+      "step": 5970
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.362657091561939e-05,
+      "loss": 0.1547,
+      "step": 5980
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 4.3606622780770004e-05,
+      "loss": 0.152,
+      "step": 5990
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 4.358667464592061e-05,
+      "loss": 0.1615,
+      "step": 6000
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 4.3566726511071215e-05,
+      "loss": 0.1519,
+      "step": 6010
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 4.354677837622183e-05,
+      "loss": 0.1844,
+      "step": 6020
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 4.352683024137243e-05,
+      "loss": 0.1562,
+      "step": 6030
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.3506882106523046e-05,
+      "loss": 0.1823,
+      "step": 6040
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.348693397167365e-05,
+      "loss": 0.1665,
+      "step": 6050
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.346698583682426e-05,
+      "loss": 0.1796,
+      "step": 6060
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 4.344703770197487e-05,
+      "loss": 0.1753,
+      "step": 6070
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 4.342708956712548e-05,
+      "loss": 0.151,
+      "step": 6080
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 4.340714143227608e-05,
+      "loss": 0.163,
+      "step": 6090
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.338719329742669e-05,
+      "loss": 0.1853,
+      "step": 6100
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 4.3367245162577305e-05,
+      "loss": 0.1569,
+      "step": 6110
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 4.334729702772791e-05,
+      "loss": 0.1639,
+      "step": 6120
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.3327348892878516e-05,
+      "loss": 0.1744,
+      "step": 6130
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.9663601704418031
+      },
+      "eval_f1": {
+        "f1": 0.9642377959067083
+      },
+      "eval_loss": 0.09443444013595581,
+      "eval_precision": {
+        "precision": 0.9653800416731579
+      },
+      "eval_recall": {
+        "recall": 0.963250134618939
+      },
+      "eval_runtime": 98.9219,
+      "eval_samples_per_second": 180.304,
+      "eval_steps_per_second": 5.641,
+      "step": 6132
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 4.330740075802913e-05,
+      "loss": 0.1373,
+      "step": 6140
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.3287452623179734e-05,
+      "loss": 0.1833,
+      "step": 6150
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 4.326750448833034e-05,
+      "loss": 0.1554,
+      "step": 6160
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 4.324755635348095e-05,
+      "loss": 0.1551,
+      "step": 6170
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 4.322760821863156e-05,
+      "loss": 0.1708,
+      "step": 6180
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 4.320766008378217e-05,
+      "loss": 0.1303,
+      "step": 6190
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 4.3187711948932776e-05,
+      "loss": 0.1771,
+      "step": 6200
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 4.316776381408338e-05,
+      "loss": 0.1407,
+      "step": 6210
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 4.3147815679233994e-05,
+      "loss": 0.1733,
+      "step": 6220
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 4.3127867544384606e-05,
+      "loss": 0.173,
+      "step": 6230
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 4.3107919409535205e-05,
+      "loss": 0.1517,
+      "step": 6240
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 4.308797127468582e-05,
+      "loss": 0.1671,
+      "step": 6250
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 4.306802313983643e-05,
+      "loss": 0.1602,
+      "step": 6260
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 4.3048075004987036e-05,
+      "loss": 0.1912,
+      "step": 6270
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 4.302812687013764e-05,
+      "loss": 0.1617,
+      "step": 6280
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 4.3008178735288254e-05,
+      "loss": 0.1535,
+      "step": 6290
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 4.298823060043886e-05,
+      "loss": 0.1574,
+      "step": 6300
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.296828246558947e-05,
+      "loss": 0.1587,
+      "step": 6310
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 4.294833433074008e-05,
+      "loss": 0.1509,
+      "step": 6320
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 4.292838619589068e-05,
+      "loss": 0.1851,
+      "step": 6330
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 4.2908438061041295e-05,
+      "loss": 0.167,
+      "step": 6340
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 4.288848992619191e-05,
+      "loss": 0.1669,
+      "step": 6350
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 4.2868541791342506e-05,
+      "loss": 0.1634,
+      "step": 6360
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 4.284859365649312e-05,
+      "loss": 0.1827,
+      "step": 6370
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 4.282864552164373e-05,
+      "loss": 0.1486,
+      "step": 6380
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 4.280869738679434e-05,
+      "loss": 0.1492,
+      "step": 6390
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 4.278874925194494e-05,
+      "loss": 0.1554,
+      "step": 6400
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 4.2768801117095555e-05,
+      "loss": 0.1812,
+      "step": 6410
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 4.274885298224616e-05,
+      "loss": 0.1765,
+      "step": 6420
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 4.272890484739677e-05,
+      "loss": 0.1653,
+      "step": 6430
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 4.270895671254738e-05,
+      "loss": 0.1772,
+      "step": 6440
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 4.2689008577697984e-05,
+      "loss": 0.1205,
+      "step": 6450
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 4.2669060442848596e-05,
+      "loss": 0.1429,
+      "step": 6460
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 4.264911230799921e-05,
+      "loss": 0.1508,
+      "step": 6470
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.262916417314981e-05,
+      "loss": 0.1559,
+      "step": 6480
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 4.260921603830042e-05,
+      "loss": 0.1448,
+      "step": 6490
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 4.258926790345103e-05,
+      "loss": 0.1579,
+      "step": 6500
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 4.256931976860164e-05,
+      "loss": 0.1582,
+      "step": 6510
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 4.2549371633752244e-05,
+      "loss": 0.1609,
+      "step": 6520
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 4.2529423498902856e-05,
+      "loss": 0.1715,
+      "step": 6530
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 4.250947536405346e-05,
+      "loss": 0.1444,
+      "step": 6540
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 4.2489527229204074e-05,
+      "loss": 0.148,
+      "step": 6550
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 4.246957909435468e-05,
+      "loss": 0.162,
+      "step": 6560
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 4.2449630959505285e-05,
+      "loss": 0.1787,
+      "step": 6570
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 4.24296828246559e-05,
+      "loss": 0.1457,
+      "step": 6580
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 4.240973468980651e-05,
+      "loss": 0.1628,
+      "step": 6590
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 4.238978655495711e-05,
+      "loss": 0.1668,
+      "step": 6600
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 4.236983842010772e-05,
+      "loss": 0.1542,
+      "step": 6610
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 4.2349890285258333e-05,
+      "loss": 0.1475,
+      "step": 6620
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.232994215040894e-05,
+      "loss": 0.1723,
+      "step": 6630
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 4.2309994015559545e-05,
+      "loss": 0.1707,
+      "step": 6640
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 4.229004588071016e-05,
+      "loss": 0.1834,
+      "step": 6650
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 4.227009774586076e-05,
+      "loss": 0.1773,
+      "step": 6660
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 4.2250149611011375e-05,
+      "loss": 0.1742,
+      "step": 6670
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 4.223020147616198e-05,
+      "loss": 0.1585,
+      "step": 6680
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 4.2210253341312586e-05,
+      "loss": 0.1482,
+      "step": 6690
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9696120206324288
+      },
+      "eval_f1": {
+        "f1": 0.9678411020044558
+      },
+      "eval_loss": 0.08285478502511978,
+      "eval_precision": {
+        "precision": 0.9677791103198417
+      },
+      "eval_recall": {
+        "recall": 0.9679224853422947
+      },
+      "eval_runtime": 98.5998,
+      "eval_samples_per_second": 180.893,
+      "eval_steps_per_second": 5.659,
+      "step": 6690
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 4.21903052064632e-05,
+      "loss": 0.1509,
+      "step": 6700
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 4.217035707161381e-05,
+      "loss": 0.1539,
+      "step": 6710
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 4.215040893676441e-05,
+      "loss": 0.1514,
+      "step": 6720
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 4.213046080191502e-05,
+      "loss": 0.1503,
+      "step": 6730
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 4.2110512667065635e-05,
+      "loss": 0.1348,
+      "step": 6740
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 4.209056453221624e-05,
+      "loss": 0.1615,
+      "step": 6750
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 4.2070616397366846e-05,
+      "loss": 0.1485,
+      "step": 6760
+    },
+    {
+      "epoch": 12.14,
+      "learning_rate": 4.205066826251746e-05,
+      "loss": 0.1467,
+      "step": 6770
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 4.2030720127668064e-05,
+      "loss": 0.1625,
+      "step": 6780
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 4.2010771992818676e-05,
+      "loss": 0.1385,
+      "step": 6790
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 4.199082385796928e-05,
+      "loss": 0.141,
+      "step": 6800
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 4.197087572311989e-05,
+      "loss": 0.146,
+      "step": 6810
+    },
+    {
+      "epoch": 12.23,
+      "learning_rate": 4.19509275882705e-05,
+      "loss": 0.1705,
+      "step": 6820
+    },
+    {
+      "epoch": 12.25,
+      "learning_rate": 4.193097945342111e-05,
+      "loss": 0.1681,
+      "step": 6830
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 4.191103131857171e-05,
+      "loss": 0.1825,
+      "step": 6840
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 4.1891083183722323e-05,
+      "loss": 0.1443,
+      "step": 6850
+    },
+    {
+      "epoch": 12.3,
+      "learning_rate": 4.1871135048872936e-05,
+      "loss": 0.1545,
+      "step": 6860
+    },
+    {
+      "epoch": 12.32,
+      "learning_rate": 4.185118691402354e-05,
+      "loss": 0.1628,
+      "step": 6870
+    },
+    {
+      "epoch": 12.34,
+      "learning_rate": 4.183123877917415e-05,
+      "loss": 0.1562,
+      "step": 6880
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 4.181129064432476e-05,
+      "loss": 0.1655,
+      "step": 6890
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 4.1791342509475365e-05,
+      "loss": 0.1434,
+      "step": 6900
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 4.177139437462597e-05,
+      "loss": 0.1807,
+      "step": 6910
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 4.175144623977658e-05,
+      "loss": 0.1416,
+      "step": 6920
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 4.173149810492719e-05,
+      "loss": 0.1541,
+      "step": 6930
+    },
+    {
+      "epoch": 12.45,
+      "learning_rate": 4.17115499700778e-05,
+      "loss": 0.1531,
+      "step": 6940
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 4.1691601835228407e-05,
+      "loss": 0.1595,
+      "step": 6950
+    },
+    {
+      "epoch": 12.48,
+      "learning_rate": 4.167165370037901e-05,
+      "loss": 0.1723,
+      "step": 6960
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 4.1651705565529625e-05,
+      "loss": 0.1389,
+      "step": 6970
+    },
+    {
+      "epoch": 12.52,
+      "learning_rate": 4.163175743068024e-05,
+      "loss": 0.1727,
+      "step": 6980
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 4.161180929583084e-05,
+      "loss": 0.1334,
+      "step": 6990
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 4.159186116098145e-05,
+      "loss": 0.1305,
+      "step": 7000
+    },
+    {
+      "epoch": 12.57,
+      "learning_rate": 4.157191302613206e-05,
+      "loss": 0.1279,
+      "step": 7010
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.1551964891282666e-05,
+      "loss": 0.1575,
+      "step": 7020
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 4.153201675643327e-05,
+      "loss": 0.1509,
+      "step": 7030
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 4.1512068621583884e-05,
+      "loss": 0.1735,
+      "step": 7040
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 4.149212048673449e-05,
+      "loss": 0.1442,
+      "step": 7050
+    },
+    {
+      "epoch": 12.66,
+      "learning_rate": 4.14721723518851e-05,
+      "loss": 0.1743,
+      "step": 7060
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.145222421703571e-05,
+      "loss": 0.1438,
+      "step": 7070
+    },
+    {
+      "epoch": 12.7,
+      "learning_rate": 4.143227608218631e-05,
+      "loss": 0.1572,
+      "step": 7080
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 4.1412327947336926e-05,
+      "loss": 0.1466,
+      "step": 7090
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 4.139237981248754e-05,
+      "loss": 0.1528,
+      "step": 7100
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 4.1372431677638144e-05,
+      "loss": 0.1478,
+      "step": 7110
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 4.135248354278875e-05,
+      "loss": 0.1664,
+      "step": 7120
+    },
+    {
+      "epoch": 12.79,
+      "learning_rate": 4.133253540793936e-05,
+      "loss": 0.1398,
+      "step": 7130
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 4.131258727308997e-05,
+      "loss": 0.1502,
+      "step": 7140
+    },
+    {
+      "epoch": 12.83,
+      "learning_rate": 4.129263913824057e-05,
+      "loss": 0.1756,
+      "step": 7150
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 4.1272691003391185e-05,
+      "loss": 0.1391,
+      "step": 7160
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 4.125274286854179e-05,
+      "loss": 0.1591,
+      "step": 7170
+    },
+    {
+      "epoch": 12.88,
+      "learning_rate": 4.12327947336924e-05,
+      "loss": 0.137,
+      "step": 7180
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 4.121284659884301e-05,
+      "loss": 0.1369,
+      "step": 7190
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 4.1192898463993615e-05,
+      "loss": 0.1359,
+      "step": 7200
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 4.117295032914423e-05,
+      "loss": 0.1587,
+      "step": 7210
+    },
+    {
+      "epoch": 12.95,
+      "learning_rate": 4.115300219429484e-05,
+      "loss": 0.1315,
+      "step": 7220
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 4.1133054059445445e-05,
+      "loss": 0.1685,
+      "step": 7230
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 4.111310592459605e-05,
+      "loss": 0.1598,
+      "step": 7240
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.967761830006728
+      },
+      "eval_f1": {
+        "f1": 0.9657620776941498
+      },
+      "eval_loss": 0.09193319082260132,
+      "eval_precision": {
+        "precision": 0.9671382733917124
+      },
+      "eval_recall": {
+        "recall": 0.9645924834051133
+      },
+      "eval_runtime": 95.698,
+      "eval_samples_per_second": 186.378,
+      "eval_steps_per_second": 5.831,
+      "step": 7247
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 4.109315778974666e-05,
+      "loss": 0.1567,
+      "step": 7250
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 4.107320965489727e-05,
+      "loss": 0.1456,
+      "step": 7260
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 4.1053261520047874e-05,
+      "loss": 0.1383,
+      "step": 7270
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 4.1033313385198486e-05,
+      "loss": 0.1416,
+      "step": 7280
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 4.101336525034909e-05,
+      "loss": 0.1572,
+      "step": 7290
+    },
+    {
+      "epoch": 13.09,
+      "learning_rate": 4.0993417115499704e-05,
+      "loss": 0.165,
+      "step": 7300
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 4.097346898065031e-05,
+      "loss": 0.1564,
+      "step": 7310
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 4.0953520845800916e-05,
+      "loss": 0.148,
+      "step": 7320
+    },
+    {
+      "epoch": 13.15,
+      "learning_rate": 4.093357271095153e-05,
+      "loss": 0.1406,
+      "step": 7330
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 4.091362457610214e-05,
+      "loss": 0.1441,
+      "step": 7340
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 4.0893676441252746e-05,
+      "loss": 0.1341,
+      "step": 7350
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 4.087372830640335e-05,
+      "loss": 0.1735,
+      "step": 7360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 4.0853780171553964e-05,
+      "loss": 0.1473,
+      "step": 7370
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 4.083383203670457e-05,
+      "loss": 0.1515,
+      "step": 7380
+    },
+    {
+      "epoch": 13.26,
+      "learning_rate": 4.0813883901855175e-05,
+      "loss": 0.133,
+      "step": 7390
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 4.079393576700579e-05,
+      "loss": 0.1275,
+      "step": 7400
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 4.077398763215639e-05,
+      "loss": 0.1491,
+      "step": 7410
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 4.0754039497307006e-05,
+      "loss": 0.1358,
+      "step": 7420
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 4.073409136245761e-05,
+      "loss": 0.1438,
+      "step": 7430
+    },
+    {
+      "epoch": 13.35,
+      "learning_rate": 4.071414322760822e-05,
+      "loss": 0.1366,
+      "step": 7440
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 4.069419509275883e-05,
+      "loss": 0.1528,
+      "step": 7450
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 4.067424695790944e-05,
+      "loss": 0.1153,
+      "step": 7460
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 4.065429882306005e-05,
+      "loss": 0.1614,
+      "step": 7470
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 4.063435068821065e-05,
+      "loss": 0.1493,
+      "step": 7480
+    },
+    {
+      "epoch": 13.43,
+      "learning_rate": 4.0614402553361265e-05,
+      "loss": 0.1667,
+      "step": 7490
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 4.059445441851187e-05,
+      "loss": 0.1531,
+      "step": 7500
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 4.0574506283662476e-05,
+      "loss": 0.1608,
+      "step": 7510
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 4.055455814881309e-05,
+      "loss": 0.16,
+      "step": 7520
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 4.0534610013963694e-05,
+      "loss": 0.1384,
+      "step": 7530
+    },
+    {
+      "epoch": 13.52,
+      "learning_rate": 4.051466187911431e-05,
+      "loss": 0.1393,
+      "step": 7540
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 4.049471374426491e-05,
+      "loss": 0.1421,
+      "step": 7550
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 4.0474765609415525e-05,
+      "loss": 0.1609,
+      "step": 7560
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 4.045481747456613e-05,
+      "loss": 0.145,
+      "step": 7570
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 4.043486933971674e-05,
+      "loss": 0.1579,
+      "step": 7580
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 4.041492120486735e-05,
+      "loss": 0.1266,
+      "step": 7590
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 4.0394973070017954e-05,
+      "loss": 0.1368,
+      "step": 7600
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 4.0375024935168566e-05,
+      "loss": 0.1229,
+      "step": 7610
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 4.035507680031917e-05,
+      "loss": 0.1703,
+      "step": 7620
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 4.033512866546978e-05,
+      "loss": 0.1543,
+      "step": 7630
+    },
+    {
+      "epoch": 13.7,
+      "learning_rate": 4.031518053062039e-05,
+      "loss": 0.1762,
+      "step": 7640
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 4.0295232395770996e-05,
+      "loss": 0.1034,
+      "step": 7650
+    },
+    {
+      "epoch": 13.74,
+      "learning_rate": 4.02752842609216e-05,
+      "loss": 0.1531,
+      "step": 7660
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 4.0255336126072214e-05,
+      "loss": 0.1451,
+      "step": 7670
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 4.0235387991222826e-05,
+      "loss": 0.146,
+      "step": 7680
+    },
+    {
+      "epoch": 13.79,
+      "learning_rate": 4.021543985637343e-05,
+      "loss": 0.1406,
+      "step": 7690
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 4.019549172152404e-05,
+      "loss": 0.1423,
+      "step": 7700
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 4.017554358667465e-05,
+      "loss": 0.1547,
+      "step": 7710
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 4.0155595451825255e-05,
+      "loss": 0.1531,
+      "step": 7720
+    },
+    {
+      "epoch": 13.87,
+      "learning_rate": 4.013564731697587e-05,
+      "loss": 0.1458,
+      "step": 7730
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 4.011569918212647e-05,
+      "loss": 0.1192,
+      "step": 7740
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 4.009575104727708e-05,
+      "loss": 0.1615,
+      "step": 7750
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 4.007580291242769e-05,
+      "loss": 0.1108,
+      "step": 7760
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 4.00558547775783e-05,
+      "loss": 0.1445,
+      "step": 7770
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 4.00359066427289e-05,
+      "loss": 0.1466,
+      "step": 7780
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 4.0015958507879515e-05,
+      "loss": 0.1465,
+      "step": 7790
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 3.999601037303013e-05,
+      "loss": 0.1614,
+      "step": 7800
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9706772819017717
+      },
+      "eval_f1": {
+        "f1": 0.968670422404098
+      },
+      "eval_loss": 0.0847558081150055,
+      "eval_precision": {
+        "precision": 0.9697786748358946
+      },
+      "eval_recall": {
+        "recall": 0.9678101937075398
+      },
+      "eval_runtime": 95.5289,
+      "eval_samples_per_second": 186.708,
+      "eval_steps_per_second": 5.841,
+      "step": 7805
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 3.997606223818073e-05,
+      "loss": 0.1499,
+      "step": 7810
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 3.995611410333134e-05,
+      "loss": 0.1341,
+      "step": 7820
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 3.993616596848195e-05,
+      "loss": 0.1476,
+      "step": 7830
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 3.9916217833632556e-05,
+      "loss": 0.1378,
+      "step": 7840
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 3.989626969878317e-05,
+      "loss": 0.1444,
+      "step": 7850
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 3.9876321563933774e-05,
+      "loss": 0.1529,
+      "step": 7860
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 3.985637342908438e-05,
+      "loss": 0.1619,
+      "step": 7870
+    },
+    {
+      "epoch": 14.13,
+      "learning_rate": 3.983642529423499e-05,
+      "loss": 0.1228,
+      "step": 7880
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 3.98164771593856e-05,
+      "loss": 0.1469,
+      "step": 7890
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 3.9796529024536204e-05,
+      "loss": 0.1302,
+      "step": 7900
+    },
+    {
+      "epoch": 14.19,
+      "learning_rate": 3.9776580889686816e-05,
+      "loss": 0.1513,
+      "step": 7910
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 3.975663275483743e-05,
+      "loss": 0.1471,
+      "step": 7920
+    },
+    {
+      "epoch": 14.22,
+      "learning_rate": 3.9736684619988034e-05,
+      "loss": 0.1152,
+      "step": 7930
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 3.971673648513864e-05,
+      "loss": 0.1431,
+      "step": 7940
+    },
+    {
+      "epoch": 14.26,
+      "learning_rate": 3.969678835028925e-05,
+      "loss": 0.1416,
+      "step": 7950
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 3.967684021543986e-05,
+      "loss": 0.1271,
+      "step": 7960
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 3.965689208059047e-05,
+      "loss": 0.16,
+      "step": 7970
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 3.9636943945741075e-05,
+      "loss": 0.1446,
+      "step": 7980
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 3.961699581089168e-05,
+      "loss": 0.1255,
+      "step": 7990
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 3.9597047676042293e-05,
+      "loss": 0.1516,
+      "step": 8000
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 3.95770995411929e-05,
+      "loss": 0.1346,
+      "step": 8010
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 3.9557151406343505e-05,
+      "loss": 0.1471,
+      "step": 8020
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 3.953720327149412e-05,
+      "loss": 0.1271,
+      "step": 8030
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 3.951725513664473e-05,
+      "loss": 0.1377,
+      "step": 8040
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 3.9497307001795335e-05,
+      "loss": 0.1511,
+      "step": 8050
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 3.947735886694594e-05,
+      "loss": 0.1446,
+      "step": 8060
+    },
+    {
+      "epoch": 14.48,
+      "learning_rate": 3.945741073209655e-05,
+      "loss": 0.1224,
+      "step": 8070
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 3.943746259724716e-05,
+      "loss": 0.1333,
+      "step": 8080
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 3.941751446239777e-05,
+      "loss": 0.1261,
+      "step": 8090
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 3.939756632754838e-05,
+      "loss": 0.1218,
+      "step": 8100
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 3.937761819269898e-05,
+      "loss": 0.1288,
+      "step": 8110
+    },
+    {
+      "epoch": 14.57,
+      "learning_rate": 3.9357670057849595e-05,
+      "loss": 0.1381,
+      "step": 8120
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 3.933772192300021e-05,
+      "loss": 0.1675,
+      "step": 8130
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.9317773788150806e-05,
+      "loss": 0.1316,
+      "step": 8140
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 3.929782565330142e-05,
+      "loss": 0.141,
+      "step": 8150
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 3.927787751845203e-05,
+      "loss": 0.1449,
+      "step": 8160
+    },
+    {
+      "epoch": 14.65,
+      "learning_rate": 3.9257929383602636e-05,
+      "loss": 0.1452,
+      "step": 8170
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 3.923798124875324e-05,
+      "loss": 0.1182,
+      "step": 8180
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 3.9218033113903854e-05,
+      "loss": 0.113,
+      "step": 8190
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 3.919808497905446e-05,
+      "loss": 0.1608,
+      "step": 8200
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 3.917813684420507e-05,
+      "loss": 0.1355,
+      "step": 8210
+    },
+    {
+      "epoch": 14.74,
+      "learning_rate": 3.915818870935568e-05,
+      "loss": 0.1525,
+      "step": 8220
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 3.9138240574506283e-05,
+      "loss": 0.1312,
+      "step": 8230
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 3.9118292439656896e-05,
+      "loss": 0.1354,
+      "step": 8240
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 3.909834430480751e-05,
+      "loss": 0.1372,
+      "step": 8250
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 3.907839616995811e-05,
+      "loss": 0.1357,
+      "step": 8260
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 3.905844803510872e-05,
+      "loss": 0.1548,
+      "step": 8270
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 3.903849990025933e-05,
+      "loss": 0.1533,
+      "step": 8280
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 3.901855176540994e-05,
+      "loss": 0.1368,
+      "step": 8290
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 3.899860363056054e-05,
+      "loss": 0.1251,
+      "step": 8300
+    },
+    {
+      "epoch": 14.91,
+      "learning_rate": 3.8978655495711155e-05,
+      "loss": 0.1482,
+      "step": 8310
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 3.895870736086176e-05,
+      "loss": 0.1466,
+      "step": 8320
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 3.893875922601237e-05,
+      "loss": 0.1545,
+      "step": 8330
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 3.891881109116298e-05,
+      "loss": 0.1422,
+      "step": 8340
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 3.8898862956313585e-05,
+      "loss": 0.1395,
+      "step": 8350
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 3.88789148214642e-05,
+      "loss": 0.1292,
+      "step": 8360
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.967873962771922
+      },
+      "eval_f1": {
+        "f1": 0.9658569332167405
+      },
+      "eval_loss": 0.09215713292360306,
+      "eval_precision": {
+        "precision": 0.9670221999602495
+      },
+      "eval_recall": {
+        "recall": 0.9648394517708
+      },
+      "eval_runtime": 95.701,
+      "eval_samples_per_second": 186.372,
+      "eval_steps_per_second": 5.831,
+      "step": 8362
+    },
+    {
+      "epoch": 15.01,
+      "learning_rate": 3.885896668661481e-05,
+      "loss": 0.1233,
+      "step": 8370
+    },
+    {
+      "epoch": 15.03,
+      "learning_rate": 3.883901855176541e-05,
+      "loss": 0.1377,
+      "step": 8380
+    },
+    {
+      "epoch": 15.05,
+      "learning_rate": 3.881907041691602e-05,
+      "loss": 0.1423,
+      "step": 8390
+    },
+    {
+      "epoch": 15.07,
+      "learning_rate": 3.879912228206663e-05,
+      "loss": 0.1214,
+      "step": 8400
+    },
+    {
+      "epoch": 15.09,
+      "learning_rate": 3.877917414721723e-05,
+      "loss": 0.1356,
+      "step": 8410
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 3.8759226012367844e-05,
+      "loss": 0.1294,
+      "step": 8420
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 3.8739277877518457e-05,
+      "loss": 0.1338,
+      "step": 8430
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 3.871932974266906e-05,
+      "loss": 0.1172,
+      "step": 8440
+    },
+    {
+      "epoch": 15.16,
+      "learning_rate": 3.869938160781967e-05,
+      "loss": 0.1272,
+      "step": 8450
+    },
+    {
+      "epoch": 15.17,
+      "learning_rate": 3.867943347297028e-05,
+      "loss": 0.1348,
+      "step": 8460
+    },
+    {
+      "epoch": 15.19,
+      "learning_rate": 3.8659485338120886e-05,
+      "loss": 0.1335,
+      "step": 8470
+    },
+    {
+      "epoch": 15.21,
+      "learning_rate": 3.86395372032715e-05,
+      "loss": 0.1382,
+      "step": 8480
+    },
+    {
+      "epoch": 15.23,
+      "learning_rate": 3.8619589068422104e-05,
+      "loss": 0.1323,
+      "step": 8490
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 3.859964093357271e-05,
+      "loss": 0.145,
+      "step": 8500
+    },
+    {
+      "epoch": 15.26,
+      "learning_rate": 3.857969279872332e-05,
+      "loss": 0.1432,
+      "step": 8510
+    },
+    {
+      "epoch": 15.28,
+      "learning_rate": 3.8559744663873934e-05,
+      "loss": 0.1568,
+      "step": 8520
+    },
+    {
+      "epoch": 15.3,
+      "learning_rate": 3.853979652902453e-05,
+      "loss": 0.1231,
+      "step": 8530
+    },
+    {
+      "epoch": 15.32,
+      "learning_rate": 3.8519848394175145e-05,
+      "loss": 0.146,
+      "step": 8540
+    },
+    {
+      "epoch": 15.34,
+      "learning_rate": 3.849990025932576e-05,
+      "loss": 0.1114,
+      "step": 8550
+    },
+    {
+      "epoch": 15.35,
+      "learning_rate": 3.847995212447636e-05,
+      "loss": 0.0976,
+      "step": 8560
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 3.846000398962697e-05,
+      "loss": 0.1269,
+      "step": 8570
+    },
+    {
+      "epoch": 15.39,
+      "learning_rate": 3.844005585477758e-05,
+      "loss": 0.1703,
+      "step": 8580
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 3.842010771992819e-05,
+      "loss": 0.1444,
+      "step": 8590
+    },
+    {
+      "epoch": 15.43,
+      "learning_rate": 3.84001595850788e-05,
+      "loss": 0.1194,
+      "step": 8600
+    },
+    {
+      "epoch": 15.44,
+      "learning_rate": 3.8380211450229405e-05,
+      "loss": 0.1645,
+      "step": 8610
+    },
+    {
+      "epoch": 15.46,
+      "learning_rate": 3.836026331538001e-05,
+      "loss": 0.1235,
+      "step": 8620
+    },
+    {
+      "epoch": 15.48,
+      "learning_rate": 3.834031518053062e-05,
+      "loss": 0.1592,
+      "step": 8630
+    },
+    {
+      "epoch": 15.5,
+      "learning_rate": 3.8320367045681235e-05,
+      "loss": 0.1244,
+      "step": 8640
+    },
+    {
+      "epoch": 15.52,
+      "learning_rate": 3.8300418910831834e-05,
+      "loss": 0.1208,
+      "step": 8650
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 3.8280470775982446e-05,
+      "loss": 0.159,
+      "step": 8660
+    },
+    {
+      "epoch": 15.55,
+      "learning_rate": 3.826052264113306e-05,
+      "loss": 0.1353,
+      "step": 8670
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 3.8240574506283664e-05,
+      "loss": 0.1368,
+      "step": 8680
+    },
+    {
+      "epoch": 15.59,
+      "learning_rate": 3.822062637143427e-05,
+      "loss": 0.133,
+      "step": 8690
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 3.820067823658488e-05,
+      "loss": 0.133,
+      "step": 8700
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 3.818073010173549e-05,
+      "loss": 0.1055,
+      "step": 8710
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 3.81607819668861e-05,
+      "loss": 0.1274,
+      "step": 8720
+    },
+    {
+      "epoch": 15.66,
+      "learning_rate": 3.8140833832036706e-05,
+      "loss": 0.1398,
+      "step": 8730
+    },
+    {
+      "epoch": 15.68,
+      "learning_rate": 3.812088569718731e-05,
+      "loss": 0.1407,
+      "step": 8740
+    },
+    {
+      "epoch": 15.7,
+      "learning_rate": 3.8100937562337924e-05,
+      "loss": 0.1103,
+      "step": 8750
+    },
+    {
+      "epoch": 15.71,
+      "learning_rate": 3.8080989427488536e-05,
+      "loss": 0.1391,
+      "step": 8760
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 3.8061041292639135e-05,
+      "loss": 0.1307,
+      "step": 8770
+    },
+    {
+      "epoch": 15.75,
+      "learning_rate": 3.804109315778975e-05,
+      "loss": 0.1219,
+      "step": 8780
+    },
+    {
+      "epoch": 15.77,
+      "learning_rate": 3.802114502294036e-05,
+      "loss": 0.1603,
+      "step": 8790
+    },
+    {
+      "epoch": 15.78,
+      "learning_rate": 3.8001196888090966e-05,
+      "loss": 0.1483,
+      "step": 8800
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 3.798124875324157e-05,
+      "loss": 0.1545,
+      "step": 8810
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 3.7961300618392184e-05,
+      "loss": 0.138,
+      "step": 8820
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 3.794135248354279e-05,
+      "loss": 0.1218,
+      "step": 8830
+    },
+    {
+      "epoch": 15.86,
+      "learning_rate": 3.79214043486934e-05,
+      "loss": 0.1481,
+      "step": 8840
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 3.790145621384401e-05,
+      "loss": 0.1387,
+      "step": 8850
+    },
+    {
+      "epoch": 15.89,
+      "learning_rate": 3.788150807899461e-05,
+      "loss": 0.1423,
+      "step": 8860
+    },
+    {
+      "epoch": 15.91,
+      "learning_rate": 3.7861559944145225e-05,
+      "loss": 0.1164,
+      "step": 8870
+    },
+    {
+      "epoch": 15.93,
+      "learning_rate": 3.784161180929584e-05,
+      "loss": 0.1513,
+      "step": 8880
+    },
+    {
+      "epoch": 15.95,
+      "learning_rate": 3.7821663674446436e-05,
+      "loss": 0.1585,
+      "step": 8890
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 3.780171553959705e-05,
+      "loss": 0.1151,
+      "step": 8900
+    },
+    {
+      "epoch": 15.98,
+      "learning_rate": 3.778176740474766e-05,
+      "loss": 0.1272,
+      "step": 8910
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 3.776181926989827e-05,
+      "loss": 0.0988,
+      "step": 8920
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": {
+        "accuracy": 0.9703408836061898
+      },
+      "eval_f1": {
+        "f1": 0.9688037947343187
+      },
+      "eval_loss": 0.08405764400959015,
+      "eval_precision": {
+        "precision": 0.9683340964462768
+      },
+      "eval_recall": {
+        "recall": 0.9693524400362661
+      },
+      "eval_runtime": 96.499,
+      "eval_samples_per_second": 184.831,
+      "eval_steps_per_second": 5.782,
+      "step": 8920
+    },
+    {
+      "epoch": 16.02,
+      "learning_rate": 3.774187113504887e-05,
+      "loss": 0.0976,
+      "step": 8930
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 3.7721923000199485e-05,
+      "loss": 0.163,
+      "step": 8940
+    },
+    {
+      "epoch": 16.05,
+      "learning_rate": 3.770197486535009e-05,
+      "loss": 0.1468,
+      "step": 8950
+    },
+    {
+      "epoch": 16.07,
+      "learning_rate": 3.76820267305007e-05,
+      "loss": 0.1222,
+      "step": 8960
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 3.766207859565131e-05,
+      "loss": 0.1322,
+      "step": 8970
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 3.7642130460801914e-05,
+      "loss": 0.151,
+      "step": 8980
+    },
+    {
+      "epoch": 16.13,
+      "learning_rate": 3.7622182325952526e-05,
+      "loss": 0.1287,
+      "step": 8990
+    },
+    {
+      "epoch": 16.14,
+      "learning_rate": 3.760223419110314e-05,
+      "loss": 0.1319,
+      "step": 9000
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 3.758228605625374e-05,
+      "loss": 0.1231,
+      "step": 9010
+    },
+    {
+      "epoch": 16.18,
+      "learning_rate": 3.756233792140435e-05,
+      "loss": 0.1463,
+      "step": 9020
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 3.754238978655496e-05,
+      "loss": 0.1263,
+      "step": 9030
+    },
+    {
+      "epoch": 16.22,
+      "learning_rate": 3.752244165170557e-05,
+      "loss": 0.138,
+      "step": 9040
+    },
+    {
+      "epoch": 16.23,
+      "learning_rate": 3.7502493516856174e-05,
+      "loss": 0.1352,
+      "step": 9050
+    },
+    {
+      "epoch": 16.25,
+      "learning_rate": 3.7482545382006786e-05,
+      "loss": 0.1268,
+      "step": 9060
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 3.746259724715739e-05,
+      "loss": 0.1274,
+      "step": 9070
+    },
+    {
+      "epoch": 16.29,
+      "learning_rate": 3.7442649112308004e-05,
+      "loss": 0.1289,
+      "step": 9080
+    },
+    {
+      "epoch": 16.3,
+      "learning_rate": 3.742270097745861e-05,
+      "loss": 0.1263,
+      "step": 9090
+    },
+    {
+      "epoch": 16.32,
+      "learning_rate": 3.7402752842609215e-05,
+      "loss": 0.139,
+      "step": 9100
+    },
+    {
+      "epoch": 16.34,
+      "learning_rate": 3.738280470775983e-05,
+      "loss": 0.1366,
+      "step": 9110
+    },
+    {
+      "epoch": 16.36,
+      "learning_rate": 3.736285657291044e-05,
+      "loss": 0.1141,
+      "step": 9120
+    },
+    {
+      "epoch": 16.38,
+      "learning_rate": 3.734290843806104e-05,
+      "loss": 0.151,
+      "step": 9130
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 3.732296030321165e-05,
+      "loss": 0.1181,
+      "step": 9140
+    },
+    {
+      "epoch": 16.41,
+      "learning_rate": 3.7303012168362264e-05,
+      "loss": 0.1182,
+      "step": 9150
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 3.728306403351286e-05,
+      "loss": 0.0934,
+      "step": 9160
+    },
+    {
+      "epoch": 16.45,
+      "learning_rate": 3.7263115898663475e-05,
+      "loss": 0.1123,
+      "step": 9170
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 3.724316776381409e-05,
+      "loss": 0.1232,
+      "step": 9180
+    },
+    {
+      "epoch": 16.48,
+      "learning_rate": 3.722321962896469e-05,
+      "loss": 0.1465,
+      "step": 9190
+    },
+    {
+      "epoch": 16.5,
+      "learning_rate": 3.72032714941153e-05,
+      "loss": 0.1241,
+      "step": 9200
+    },
+    {
+      "epoch": 16.52,
+      "learning_rate": 3.718332335926591e-05,
+      "loss": 0.1229,
+      "step": 9210
+    },
+    {
+      "epoch": 16.54,
+      "learning_rate": 3.7163375224416516e-05,
+      "loss": 0.1311,
+      "step": 9220
+    },
+    {
+      "epoch": 16.56,
+      "learning_rate": 3.714342708956713e-05,
+      "loss": 0.1056,
+      "step": 9230
+    },
+    {
+      "epoch": 16.57,
+      "learning_rate": 3.7123478954717734e-05,
+      "loss": 0.1462,
+      "step": 9240
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 3.710353081986834e-05,
+      "loss": 0.1137,
+      "step": 9250
+    },
+    {
+      "epoch": 16.61,
+      "learning_rate": 3.708358268501895e-05,
+      "loss": 0.1463,
+      "step": 9260
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 3.7063634550169565e-05,
+      "loss": 0.1163,
+      "step": 9270
+    },
+    {
+      "epoch": 16.65,
+      "learning_rate": 3.7043686415320164e-05,
+      "loss": 0.1181,
+      "step": 9280
+    },
+    {
+      "epoch": 16.66,
+      "learning_rate": 3.7023738280470776e-05,
+      "loss": 0.1237,
+      "step": 9290
+    },
+    {
+      "epoch": 16.68,
+      "learning_rate": 3.700379014562139e-05,
+      "loss": 0.1316,
+      "step": 9300
+    },
+    {
+      "epoch": 16.7,
+      "learning_rate": 3.6983842010771994e-05,
+      "loss": 0.1359,
+      "step": 9310
+    },
+    {
+      "epoch": 16.72,
+      "learning_rate": 3.69638938759226e-05,
+      "loss": 0.1258,
+      "step": 9320
+    },
+    {
+      "epoch": 16.74,
+      "learning_rate": 3.694394574107321e-05,
+      "loss": 0.1433,
+      "step": 9330
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 3.692399760622382e-05,
+      "loss": 0.1082,
+      "step": 9340
+    },
+    {
+      "epoch": 16.77,
+      "learning_rate": 3.690404947137443e-05,
+      "loss": 0.1608,
+      "step": 9350
+    },
+    {
+      "epoch": 16.79,
+      "learning_rate": 3.6884101336525035e-05,
+      "loss": 0.1269,
+      "step": 9360
+    },
+    {
+      "epoch": 16.81,
+      "learning_rate": 3.686415320167564e-05,
+      "loss": 0.1603,
+      "step": 9370
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 3.6844205066826253e-05,
+      "loss": 0.1275,
+      "step": 9380
+    },
+    {
+      "epoch": 16.84,
+      "learning_rate": 3.6824256931976866e-05,
+      "loss": 0.1168,
+      "step": 9390
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 3.6804308797127465e-05,
+      "loss": 0.1231,
+      "step": 9400
+    },
+    {
+      "epoch": 16.88,
+      "learning_rate": 3.678436066227808e-05,
+      "loss": 0.1384,
+      "step": 9410
+    },
+    {
+      "epoch": 16.9,
+      "learning_rate": 3.676441252742869e-05,
+      "loss": 0.1272,
+      "step": 9420
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 3.6744464392579295e-05,
+      "loss": 0.1292,
+      "step": 9430
+    },
+    {
+      "epoch": 16.93,
+      "learning_rate": 3.67245162577299e-05,
+      "loss": 0.1193,
+      "step": 9440
+    },
+    {
+      "epoch": 16.95,
+      "learning_rate": 3.670456812288051e-05,
+      "loss": 0.1218,
+      "step": 9450
+    },
+    {
+      "epoch": 16.97,
+      "learning_rate": 3.668461998803112e-05,
+      "loss": 0.1277,
+      "step": 9460
+    },
+    {
+      "epoch": 16.99,
+      "learning_rate": 3.666467185318173e-05,
+      "loss": 0.1062,
+      "step": 9470
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": {
+        "accuracy": 0.9719107423189056
+      },
+      "eval_f1": {
+        "f1": 0.970393669051259
+      },
+      "eval_loss": 0.08178497105836868,
+      "eval_precision": {
+        "precision": 0.9706603351073347
+      },
+      "eval_recall": {
+        "recall": 0.970195220462345
+      },
+      "eval_runtime": 95.8143,
+      "eval_samples_per_second": 186.152,
+      "eval_steps_per_second": 5.824,
+      "step": 9477
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 3.664472371833234e-05,
+      "loss": 0.1213,
+      "step": 9480
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 3.662477558348294e-05,
+      "loss": 0.1373,
+      "step": 9490
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 3.6604827448633555e-05,
+      "loss": 0.1706,
+      "step": 9500
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 3.658487931378417e-05,
+      "loss": 0.1133,
+      "step": 9510
+    },
+    {
+      "epoch": 17.08,
+      "learning_rate": 3.656493117893477e-05,
+      "loss": 0.0913,
+      "step": 9520
+    },
+    {
+      "epoch": 17.09,
+      "learning_rate": 3.654498304408538e-05,
+      "loss": 0.1227,
+      "step": 9530
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 3.652503490923599e-05,
+      "loss": 0.1178,
+      "step": 9540
+    },
+    {
+      "epoch": 17.13,
+      "learning_rate": 3.6505086774386596e-05,
+      "loss": 0.1212,
+      "step": 9550
+    },
+    {
+      "epoch": 17.15,
+      "learning_rate": 3.64851386395372e-05,
+      "loss": 0.1015,
+      "step": 9560
+    },
+    {
+      "epoch": 17.17,
+      "learning_rate": 3.6465190504687814e-05,
+      "loss": 0.1223,
+      "step": 9570
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 3.644524236983842e-05,
+      "loss": 0.1283,
+      "step": 9580
+    },
+    {
+      "epoch": 17.2,
+      "learning_rate": 3.642529423498903e-05,
+      "loss": 0.1264,
+      "step": 9590
+    },
+    {
+      "epoch": 17.22,
+      "learning_rate": 3.640534610013964e-05,
+      "loss": 0.14,
+      "step": 9600
+    },
+    {
+      "epoch": 17.24,
+      "learning_rate": 3.6385397965290243e-05,
+      "loss": 0.1286,
+      "step": 9610
+    },
+    {
+      "epoch": 17.26,
+      "learning_rate": 3.6365449830440856e-05,
+      "loss": 0.1219,
+      "step": 9620
+    },
+    {
+      "epoch": 17.27,
+      "learning_rate": 3.634550169559147e-05,
+      "loss": 0.1309,
+      "step": 9630
+    },
+    {
+      "epoch": 17.29,
+      "learning_rate": 3.6325553560742074e-05,
+      "loss": 0.124,
+      "step": 9640
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 3.630560542589268e-05,
+      "loss": 0.1238,
+      "step": 9650
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 3.628565729104329e-05,
+      "loss": 0.1159,
+      "step": 9660
+    },
+    {
+      "epoch": 17.35,
+      "learning_rate": 3.62657091561939e-05,
+      "loss": 0.1226,
+      "step": 9670
+    },
+    {
+      "epoch": 17.36,
+      "learning_rate": 3.62457610213445e-05,
+      "loss": 0.148,
+      "step": 9680
+    },
+    {
+      "epoch": 17.38,
+      "learning_rate": 3.6225812886495115e-05,
+      "loss": 0.1352,
+      "step": 9690
+    },
+    {
+      "epoch": 17.4,
+      "learning_rate": 3.620586475164572e-05,
+      "loss": 0.1102,
+      "step": 9700
+    },
+    {
+      "epoch": 17.42,
+      "learning_rate": 3.618591661679633e-05,
+      "loss": 0.1201,
+      "step": 9710
+    },
+    {
+      "epoch": 17.43,
+      "learning_rate": 3.616596848194694e-05,
+      "loss": 0.131,
+      "step": 9720
+    },
+    {
+      "epoch": 17.45,
+      "learning_rate": 3.6146020347097545e-05,
+      "loss": 0.118,
+      "step": 9730
+    },
+    {
+      "epoch": 17.47,
+      "learning_rate": 3.612607221224816e-05,
+      "loss": 0.1221,
+      "step": 9740
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 3.610612407739877e-05,
+      "loss": 0.126,
+      "step": 9750
+    },
+    {
+      "epoch": 17.51,
+      "learning_rate": 3.6086175942549375e-05,
+      "loss": 0.1265,
+      "step": 9760
+    },
+    {
+      "epoch": 17.52,
+      "learning_rate": 3.606622780769998e-05,
+      "loss": 0.1301,
+      "step": 9770
+    },
+    {
+      "epoch": 17.54,
+      "learning_rate": 3.604627967285059e-05,
+      "loss": 0.1212,
+      "step": 9780
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 3.60263315380012e-05,
+      "loss": 0.1337,
+      "step": 9790
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 3.6006383403151804e-05,
+      "loss": 0.1325,
+      "step": 9800
+    },
+    {
+      "epoch": 17.6,
+      "learning_rate": 3.5986435268302417e-05,
+      "loss": 0.1252,
+      "step": 9810
+    },
+    {
+      "epoch": 17.61,
+      "learning_rate": 3.596648713345302e-05,
+      "loss": 0.1428,
+      "step": 9820
+    },
+    {
+      "epoch": 17.63,
+      "learning_rate": 3.5946538998603635e-05,
+      "loss": 0.1182,
+      "step": 9830
+    },
+    {
+      "epoch": 17.65,
+      "learning_rate": 3.592659086375424e-05,
+      "loss": 0.1193,
+      "step": 9840
+    },
+    {
+      "epoch": 17.67,
+      "learning_rate": 3.5906642728904846e-05,
+      "loss": 0.1111,
+      "step": 9850
+    },
+    {
+      "epoch": 17.69,
+      "learning_rate": 3.588669459405546e-05,
+      "loss": 0.1123,
+      "step": 9860
+    },
+    {
+      "epoch": 17.7,
+      "learning_rate": 3.586674645920607e-05,
+      "loss": 0.1246,
+      "step": 9870
+    },
+    {
+      "epoch": 17.72,
+      "learning_rate": 3.5846798324356676e-05,
+      "loss": 0.1229,
+      "step": 9880
+    },
+    {
+      "epoch": 17.74,
+      "learning_rate": 3.582685018950728e-05,
+      "loss": 0.1363,
+      "step": 9890
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 3.5806902054657894e-05,
+      "loss": 0.1259,
+      "step": 9900
+    },
+    {
+      "epoch": 17.78,
+      "learning_rate": 3.57869539198085e-05,
+      "loss": 0.0848,
+      "step": 9910
+    },
+    {
+      "epoch": 17.79,
+      "learning_rate": 3.5767005784959105e-05,
+      "loss": 0.0918,
+      "step": 9920
+    },
+    {
+      "epoch": 17.81,
+      "learning_rate": 3.574705765010972e-05,
+      "loss": 0.1319,
+      "step": 9930
+    },
+    {
+      "epoch": 17.83,
+      "learning_rate": 3.572710951526032e-05,
+      "loss": 0.1291,
+      "step": 9940
+    },
+    {
+      "epoch": 17.85,
+      "learning_rate": 3.570716138041093e-05,
+      "loss": 0.1178,
+      "step": 9950
+    },
+    {
+      "epoch": 17.87,
+      "learning_rate": 3.568721324556154e-05,
+      "loss": 0.1135,
+      "step": 9960
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 3.566726511071215e-05,
+      "loss": 0.1298,
+      "step": 9970
+    },
+    {
+      "epoch": 17.9,
+      "learning_rate": 3.564731697586276e-05,
+      "loss": 0.1109,
+      "step": 9980
+    },
+    {
+      "epoch": 17.92,
+      "learning_rate": 3.5627368841013365e-05,
+      "loss": 0.1326,
+      "step": 9990
+    },
+    {
+      "epoch": 17.94,
+      "learning_rate": 3.560742070616398e-05,
+      "loss": 0.1062,
+      "step": 10000
+    },
+    {
+      "epoch": 17.96,
+      "learning_rate": 3.558747257131458e-05,
+      "loss": 0.1356,
+      "step": 10010
+    },
+    {
+      "epoch": 17.97,
+      "learning_rate": 3.5567524436465195e-05,
+      "loss": 0.1279,
+      "step": 10020
+    },
+    {
+      "epoch": 17.99,
+      "learning_rate": 3.55475763016158e-05,
+      "loss": 0.1291,
+      "step": 10030
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": {
+        "accuracy": 0.9703969499887867
+      },
+      "eval_f1": {
+        "f1": 0.9687676965848208
+      },
+      "eval_loss": 0.08788565546274185,
+      "eval_precision": {
+        "precision": 0.9686849139922615
+      },
+      "eval_recall": {
+        "recall": 0.9690100916019242
+      },
+      "eval_runtime": 96.3636,
+      "eval_samples_per_second": 185.091,
+      "eval_steps_per_second": 5.791,
+      "step": 10035
+    },
+    {
+      "epoch": 18.01,
+      "learning_rate": 3.5527628166766406e-05,
+      "loss": 0.1297,
+      "step": 10040
+    },
+    {
+      "epoch": 18.03,
+      "learning_rate": 3.550768003191702e-05,
+      "loss": 0.1155,
+      "step": 10050
+    },
+    {
+      "epoch": 18.04,
+      "learning_rate": 3.5487731897067624e-05,
+      "loss": 0.0892,
+      "step": 10060
+    },
+    {
+      "epoch": 18.06,
+      "learning_rate": 3.546778376221823e-05,
+      "loss": 0.1048,
+      "step": 10070
+    },
+    {
+      "epoch": 18.08,
+      "learning_rate": 3.544783562736884e-05,
+      "loss": 0.1328,
+      "step": 10080
+    },
+    {
+      "epoch": 18.1,
+      "learning_rate": 3.5427887492519455e-05,
+      "loss": 0.1087,
+      "step": 10090
+    },
+    {
+      "epoch": 18.12,
+      "learning_rate": 3.540793935767006e-05,
+      "loss": 0.0989,
+      "step": 10100
+    },
+    {
+      "epoch": 18.13,
+      "learning_rate": 3.5387991222820666e-05,
+      "loss": 0.1456,
+      "step": 10110
+    },
+    {
+      "epoch": 18.15,
+      "learning_rate": 3.536804308797128e-05,
+      "loss": 0.1244,
+      "step": 10120
+    },
+    {
+      "epoch": 18.17,
+      "learning_rate": 3.5348094953121884e-05,
+      "loss": 0.1182,
+      "step": 10130
+    },
+    {
+      "epoch": 18.19,
+      "learning_rate": 3.5328146818272496e-05,
+      "loss": 0.126,
+      "step": 10140
+    },
+    {
+      "epoch": 18.21,
+      "learning_rate": 3.53081986834231e-05,
+      "loss": 0.105,
+      "step": 10150
+    },
+    {
+      "epoch": 18.22,
+      "learning_rate": 3.528825054857371e-05,
+      "loss": 0.1367,
+      "step": 10160
+    },
+    {
+      "epoch": 18.24,
+      "learning_rate": 3.526830241372432e-05,
+      "loss": 0.1155,
+      "step": 10170
+    },
+    {
+      "epoch": 18.26,
+      "learning_rate": 3.5248354278874926e-05,
+      "loss": 0.1011,
+      "step": 10180
+    },
+    {
+      "epoch": 18.28,
+      "learning_rate": 3.522840614402553e-05,
+      "loss": 0.1189,
+      "step": 10190
+    },
+    {
+      "epoch": 18.3,
+      "learning_rate": 3.5208458009176144e-05,
+      "loss": 0.1124,
+      "step": 10200
+    },
+    {
+      "epoch": 18.31,
+      "learning_rate": 3.5188509874326756e-05,
+      "loss": 0.105,
+      "step": 10210
+    },
+    {
+      "epoch": 18.33,
+      "learning_rate": 3.516856173947736e-05,
+      "loss": 0.1166,
+      "step": 10220
+    },
+    {
+      "epoch": 18.35,
+      "learning_rate": 3.514861360462797e-05,
+      "loss": 0.1119,
+      "step": 10230
+    },
+    {
+      "epoch": 18.37,
+      "learning_rate": 3.512866546977858e-05,
+      "loss": 0.1337,
+      "step": 10240
+    },
+    {
+      "epoch": 18.39,
+      "learning_rate": 3.5108717334929185e-05,
+      "loss": 0.1168,
+      "step": 10250
+    },
+    {
+      "epoch": 18.4,
+      "learning_rate": 3.50887692000798e-05,
+      "loss": 0.1152,
+      "step": 10260
+    },
+    {
+      "epoch": 18.42,
+      "learning_rate": 3.50688210652304e-05,
+      "loss": 0.1319,
+      "step": 10270
+    },
+    {
+      "epoch": 18.44,
+      "learning_rate": 3.504887293038101e-05,
+      "loss": 0.1266,
+      "step": 10280
+    },
+    {
+      "epoch": 18.46,
+      "learning_rate": 3.502892479553162e-05,
+      "loss": 0.1214,
+      "step": 10290
+    },
+    {
+      "epoch": 18.48,
+      "learning_rate": 3.500897666068223e-05,
+      "loss": 0.1489,
+      "step": 10300
+    },
+    {
+      "epoch": 18.49,
+      "learning_rate": 3.498902852583283e-05,
+      "loss": 0.1094,
+      "step": 10310
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 3.4969080390983445e-05,
+      "loss": 0.102,
+      "step": 10320
+    },
+    {
+      "epoch": 18.53,
+      "learning_rate": 3.494913225613406e-05,
+      "loss": 0.1277,
+      "step": 10330
+    },
+    {
+      "epoch": 18.55,
+      "learning_rate": 3.492918412128466e-05,
+      "loss": 0.1265,
+      "step": 10340
+    },
+    {
+      "epoch": 18.57,
+      "learning_rate": 3.490923598643527e-05,
+      "loss": 0.1156,
+      "step": 10350
+    },
+    {
+      "epoch": 18.58,
+      "learning_rate": 3.488928785158588e-05,
+      "loss": 0.1349,
+      "step": 10360
+    },
+    {
+      "epoch": 18.6,
+      "learning_rate": 3.4869339716736486e-05,
+      "loss": 0.1305,
+      "step": 10370
+    },
+    {
+      "epoch": 18.62,
+      "learning_rate": 3.48493915818871e-05,
+      "loss": 0.1159,
+      "step": 10380
+    },
+    {
+      "epoch": 18.64,
+      "learning_rate": 3.4829443447037704e-05,
+      "loss": 0.1168,
+      "step": 10390
+    },
+    {
+      "epoch": 18.65,
+      "learning_rate": 3.480949531218831e-05,
+      "loss": 0.1233,
+      "step": 10400
+    },
+    {
+      "epoch": 18.67,
+      "learning_rate": 3.478954717733892e-05,
+      "loss": 0.1179,
+      "step": 10410
+    },
+    {
+      "epoch": 18.69,
+      "learning_rate": 3.476959904248953e-05,
+      "loss": 0.1384,
+      "step": 10420
+    },
+    {
+      "epoch": 18.71,
+      "learning_rate": 3.4749650907640134e-05,
+      "loss": 0.1282,
+      "step": 10430
+    },
+    {
+      "epoch": 18.73,
+      "learning_rate": 3.4729702772790746e-05,
+      "loss": 0.1296,
+      "step": 10440
+    },
+    {
+      "epoch": 18.74,
+      "learning_rate": 3.470975463794136e-05,
+      "loss": 0.1343,
+      "step": 10450
+    },
+    {
+      "epoch": 18.76,
+      "learning_rate": 3.4689806503091964e-05,
+      "loss": 0.1446,
+      "step": 10460
+    },
+    {
+      "epoch": 18.78,
+      "learning_rate": 3.466985836824257e-05,
+      "loss": 0.1107,
+      "step": 10470
+    },
+    {
+      "epoch": 18.8,
+      "learning_rate": 3.464991023339318e-05,
+      "loss": 0.1009,
+      "step": 10480
+    },
+    {
+      "epoch": 18.82,
+      "learning_rate": 3.462996209854379e-05,
+      "loss": 0.1396,
+      "step": 10490
+    },
+    {
+      "epoch": 18.83,
+      "learning_rate": 3.46100139636944e-05,
+      "loss": 0.1205,
+      "step": 10500
+    },
+    {
+      "epoch": 18.85,
+      "learning_rate": 3.4590065828845006e-05,
+      "loss": 0.1426,
+      "step": 10510
+    },
+    {
+      "epoch": 18.87,
+      "learning_rate": 3.457011769399561e-05,
+      "loss": 0.1231,
+      "step": 10520
+    },
+    {
+      "epoch": 18.89,
+      "learning_rate": 3.4550169559146224e-05,
+      "loss": 0.1107,
+      "step": 10530
+    },
+    {
+      "epoch": 18.91,
+      "learning_rate": 3.453022142429683e-05,
+      "loss": 0.1136,
+      "step": 10540
+    },
+    {
+      "epoch": 18.92,
+      "learning_rate": 3.4510273289447435e-05,
+      "loss": 0.1152,
+      "step": 10550
+    },
+    {
+      "epoch": 18.94,
+      "learning_rate": 3.449032515459805e-05,
+      "loss": 0.1121,
+      "step": 10560
+    },
+    {
+      "epoch": 18.96,
+      "learning_rate": 3.447037701974866e-05,
+      "loss": 0.103,
+      "step": 10570
+    },
+    {
+      "epoch": 18.98,
+      "learning_rate": 3.4450428884899265e-05,
+      "loss": 0.1077,
+      "step": 10580
+    },
+    {
+      "epoch": 19.0,
+      "learning_rate": 3.443048075004987e-05,
+      "loss": 0.1253,
+      "step": 10590
+    },
+    {
+      "epoch": 19.0,
+      "eval_accuracy": {
+        "accuracy": 0.9731442027360395
+      },
+      "eval_f1": {
+        "f1": 0.9716105864810595
+      },
+      "eval_loss": 0.07871522009372711,
+      "eval_precision": {
+        "precision": 0.9721612839062381
+      },
+      "eval_recall": {
+        "recall": 0.9711479618814062
+      },
+      "eval_runtime": 97.4964,
+      "eval_samples_per_second": 182.94,
+      "eval_steps_per_second": 5.723,
+      "step": 10592
+    },
+    {
+      "epoch": 19.01,
+      "learning_rate": 3.441053261520048e-05,
+      "loss": 0.1217,
+      "step": 10600
+    },
+    {
+      "epoch": 19.03,
+      "learning_rate": 3.439058448035109e-05,
+      "loss": 0.118,
+      "step": 10610
+    },
+    {
+      "epoch": 19.05,
+      "learning_rate": 3.43706363455017e-05,
+      "loss": 0.1315,
+      "step": 10620
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 3.435068821065231e-05,
+      "loss": 0.1029,
+      "step": 10630
+    },
+    {
+      "epoch": 19.09,
+      "learning_rate": 3.433074007580291e-05,
+      "loss": 0.1138,
+      "step": 10640
+    },
+    {
+      "epoch": 19.1,
+      "learning_rate": 3.4310791940953525e-05,
+      "loss": 0.1146,
+      "step": 10650
+    },
+    {
+      "epoch": 19.12,
+      "learning_rate": 3.429084380610413e-05,
+      "loss": 0.1324,
+      "step": 10660
+    },
+    {
+      "epoch": 19.14,
+      "learning_rate": 3.4270895671254736e-05,
+      "loss": 0.114,
+      "step": 10670
+    },
+    {
+      "epoch": 19.16,
+      "learning_rate": 3.425094753640535e-05,
+      "loss": 0.1205,
+      "step": 10680
+    },
+    {
+      "epoch": 19.17,
+      "learning_rate": 3.423099940155596e-05,
+      "loss": 0.1122,
+      "step": 10690
+    },
+    {
+      "epoch": 19.19,
+      "learning_rate": 3.421105126670656e-05,
+      "loss": 0.1213,
+      "step": 10700
+    },
+    {
+      "epoch": 19.21,
+      "learning_rate": 3.419110313185717e-05,
+      "loss": 0.1276,
+      "step": 10710
+    },
+    {
+      "epoch": 19.23,
+      "learning_rate": 3.4171154997007784e-05,
+      "loss": 0.0995,
+      "step": 10720
+    },
+    {
+      "epoch": 19.25,
+      "learning_rate": 3.415120686215839e-05,
+      "loss": 0.1114,
+      "step": 10730
+    },
+    {
+      "epoch": 19.26,
+      "learning_rate": 3.4131258727308995e-05,
+      "loss": 0.1356,
+      "step": 10740
+    },
+    {
+      "epoch": 19.28,
+      "learning_rate": 3.411131059245961e-05,
+      "loss": 0.133,
+      "step": 10750
+    },
+    {
+      "epoch": 19.3,
+      "learning_rate": 3.4091362457610213e-05,
+      "loss": 0.1009,
+      "step": 10760
+    },
+    {
+      "epoch": 19.32,
+      "learning_rate": 3.4071414322760826e-05,
+      "loss": 0.136,
+      "step": 10770
+    },
+    {
+      "epoch": 19.34,
+      "learning_rate": 3.405146618791143e-05,
+      "loss": 0.1111,
+      "step": 10780
+    },
+    {
+      "epoch": 19.35,
+      "learning_rate": 3.403151805306204e-05,
+      "loss": 0.1176,
+      "step": 10790
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.401156991821265e-05,
+      "loss": 0.1128,
+      "step": 10800
+    },
+    {
+      "epoch": 19.39,
+      "learning_rate": 3.399162178336326e-05,
+      "loss": 0.0974,
+      "step": 10810
+    },
+    {
+      "epoch": 19.41,
+      "learning_rate": 3.397167364851386e-05,
+      "loss": 0.1043,
+      "step": 10820
+    },
+    {
+      "epoch": 19.43,
+      "learning_rate": 3.395172551366447e-05,
+      "loss": 0.128,
+      "step": 10830
+    },
+    {
+      "epoch": 19.44,
+      "learning_rate": 3.3931777378815085e-05,
+      "loss": 0.1245,
+      "step": 10840
+    },
+    {
+      "epoch": 19.46,
+      "learning_rate": 3.391182924396569e-05,
+      "loss": 0.1144,
+      "step": 10850
+    },
+    {
+      "epoch": 19.48,
+      "learning_rate": 3.38918811091163e-05,
+      "loss": 0.1044,
+      "step": 10860
+    },
+    {
+      "epoch": 19.5,
+      "learning_rate": 3.387193297426691e-05,
+      "loss": 0.1005,
+      "step": 10870
+    },
+    {
+      "epoch": 19.52,
+      "learning_rate": 3.3851984839417515e-05,
+      "loss": 0.1083,
+      "step": 10880
+    },
+    {
+      "epoch": 19.53,
+      "learning_rate": 3.383203670456813e-05,
+      "loss": 0.129,
+      "step": 10890
+    },
+    {
+      "epoch": 19.55,
+      "learning_rate": 3.381208856971873e-05,
+      "loss": 0.1093,
+      "step": 10900
+    },
+    {
+      "epoch": 19.57,
+      "learning_rate": 3.379214043486934e-05,
+      "loss": 0.1305,
+      "step": 10910
+    },
+    {
+      "epoch": 19.59,
+      "learning_rate": 3.377219230001995e-05,
+      "loss": 0.1071,
+      "step": 10920
+    },
+    {
+      "epoch": 19.61,
+      "learning_rate": 3.375224416517056e-05,
+      "loss": 0.0956,
+      "step": 10930
+    },
+    {
+      "epoch": 19.62,
+      "learning_rate": 3.373229603032116e-05,
+      "loss": 0.1515,
+      "step": 10940
+    },
+    {
+      "epoch": 19.64,
+      "learning_rate": 3.3712347895471774e-05,
+      "loss": 0.1128,
+      "step": 10950
+    },
+    {
+      "epoch": 19.66,
+      "learning_rate": 3.3692399760622387e-05,
+      "loss": 0.1096,
+      "step": 10960
+    },
+    {
+      "epoch": 19.68,
+      "learning_rate": 3.367245162577299e-05,
+      "loss": 0.1066,
+      "step": 10970
+    },
+    {
+      "epoch": 19.7,
+      "learning_rate": 3.36525034909236e-05,
+      "loss": 0.1129,
+      "step": 10980
+    },
+    {
+      "epoch": 19.71,
+      "learning_rate": 3.363255535607421e-05,
+      "loss": 0.1279,
+      "step": 10990
+    },
+    {
+      "epoch": 19.73,
+      "learning_rate": 3.3612607221224816e-05,
+      "loss": 0.1392,
+      "step": 11000
+    },
+    {
+      "epoch": 19.75,
+      "learning_rate": 3.359265908637543e-05,
+      "loss": 0.1123,
+      "step": 11010
+    },
+    {
+      "epoch": 19.77,
+      "learning_rate": 3.3572710951526034e-05,
+      "loss": 0.1239,
+      "step": 11020
+    },
+    {
+      "epoch": 19.78,
+      "learning_rate": 3.355276281667664e-05,
+      "loss": 0.1262,
+      "step": 11030
+    },
+    {
+      "epoch": 19.8,
+      "learning_rate": 3.353281468182725e-05,
+      "loss": 0.0987,
+      "step": 11040
+    },
+    {
+      "epoch": 19.82,
+      "learning_rate": 3.3512866546977864e-05,
+      "loss": 0.1146,
+      "step": 11050
+    },
+    {
+      "epoch": 19.84,
+      "learning_rate": 3.349291841212846e-05,
+      "loss": 0.1094,
+      "step": 11060
+    },
+    {
+      "epoch": 19.86,
+      "learning_rate": 3.3472970277279075e-05,
+      "loss": 0.118,
+      "step": 11070
+    },
+    {
+      "epoch": 19.87,
+      "learning_rate": 3.345302214242969e-05,
+      "loss": 0.1014,
+      "step": 11080
+    },
+    {
+      "epoch": 19.89,
+      "learning_rate": 3.343307400758029e-05,
+      "loss": 0.1186,
+      "step": 11090
+    },
+    {
+      "epoch": 19.91,
+      "learning_rate": 3.34131258727309e-05,
+      "loss": 0.1255,
+      "step": 11100
+    },
+    {
+      "epoch": 19.93,
+      "learning_rate": 3.339317773788151e-05,
+      "loss": 0.1233,
+      "step": 11110
+    },
+    {
+      "epoch": 19.95,
+      "learning_rate": 3.337322960303212e-05,
+      "loss": 0.105,
+      "step": 11120
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 3.335328146818273e-05,
+      "loss": 0.127,
+      "step": 11130
+    },
+    {
+      "epoch": 19.98,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.1007,
+      "step": 11140
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 3.331338519848394e-05,
+      "loss": 0.1473,
+      "step": 11150
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": {
+        "accuracy": 0.9723592733796815
+      },
+      "eval_f1": {
+        "f1": 0.9711380216113251
+      },
+      "eval_loss": 0.08163481950759888,
+      "eval_precision": {
+        "precision": 0.9709134522828391
+      },
+      "eval_recall": {
+        "recall": 0.9715130547673007
+      },
+      "eval_runtime": 163.6284,
+      "eval_samples_per_second": 109.003,
+      "eval_steps_per_second": 3.41,
+      "step": 11150
+    },
+    {
+      "epoch": 20.02,
+      "learning_rate": 3.329343706363455e-05,
+      "loss": 0.104,
+      "step": 11160
+    },
+    {
+      "epoch": 20.04,
+      "learning_rate": 3.3273488928785165e-05,
+      "loss": 0.1147,
+      "step": 11170
+    },
+    {
+      "epoch": 20.05,
+      "learning_rate": 3.3253540793935764e-05,
+      "loss": 0.1285,
+      "step": 11180
+    },
+    {
+      "epoch": 20.07,
+      "learning_rate": 3.3233592659086377e-05,
+      "loss": 0.1098,
+      "step": 11190
+    },
+    {
+      "epoch": 20.09,
+      "learning_rate": 3.321364452423699e-05,
+      "loss": 0.1277,
+      "step": 11200
+    },
+    {
+      "epoch": 20.11,
+      "learning_rate": 3.3193696389387595e-05,
+      "loss": 0.132,
+      "step": 11210
+    },
+    {
+      "epoch": 20.13,
+      "learning_rate": 3.31737482545382e-05,
+      "loss": 0.1003,
+      "step": 11220
+    },
+    {
+      "epoch": 20.14,
+      "learning_rate": 3.315380011968881e-05,
+      "loss": 0.0991,
+      "step": 11230
+    },
+    {
+      "epoch": 20.16,
+      "learning_rate": 3.313385198483942e-05,
+      "loss": 0.1148,
+      "step": 11240
+    },
+    {
+      "epoch": 20.18,
+      "learning_rate": 3.311390384999003e-05,
+      "loss": 0.1156,
+      "step": 11250
+    },
+    {
+      "epoch": 20.2,
+      "learning_rate": 3.3093955715140636e-05,
+      "loss": 0.1136,
+      "step": 11260
+    },
+    {
+      "epoch": 20.22,
+      "learning_rate": 3.307400758029124e-05,
+      "loss": 0.1101,
+      "step": 11270
+    },
+    {
+      "epoch": 20.23,
+      "learning_rate": 3.3054059445441854e-05,
+      "loss": 0.112,
+      "step": 11280
+    },
+    {
+      "epoch": 20.25,
+      "learning_rate": 3.3034111310592466e-05,
+      "loss": 0.1289,
+      "step": 11290
+    },
+    {
+      "epoch": 20.27,
+      "learning_rate": 3.3014163175743065e-05,
+      "loss": 0.122,
+      "step": 11300
+    },
+    {
+      "epoch": 20.29,
+      "learning_rate": 3.299421504089368e-05,
+      "loss": 0.1383,
+      "step": 11310
+    },
+    {
+      "epoch": 20.3,
+      "learning_rate": 3.297426690604429e-05,
+      "loss": 0.1367,
+      "step": 11320
+    },
+    {
+      "epoch": 20.32,
+      "learning_rate": 3.2954318771194896e-05,
+      "loss": 0.1206,
+      "step": 11330
+    },
+    {
+      "epoch": 20.34,
+      "learning_rate": 3.29343706363455e-05,
+      "loss": 0.11,
+      "step": 11340
+    },
+    {
+      "epoch": 20.36,
+      "learning_rate": 3.2914422501496114e-05,
+      "loss": 0.1052,
+      "step": 11350
+    },
+    {
+      "epoch": 20.38,
+      "learning_rate": 3.289447436664672e-05,
+      "loss": 0.1011,
+      "step": 11360
+    },
+    {
+      "epoch": 20.39,
+      "learning_rate": 3.287452623179733e-05,
+      "loss": 0.108,
+      "step": 11370
+    },
+    {
+      "epoch": 20.41,
+      "learning_rate": 3.285457809694794e-05,
+      "loss": 0.1126,
+      "step": 11380
+    },
+    {
+      "epoch": 20.43,
+      "learning_rate": 3.283462996209854e-05,
+      "loss": 0.0932,
+      "step": 11390
+    },
+    {
+      "epoch": 20.45,
+      "learning_rate": 3.2814681827249155e-05,
+      "loss": 0.108,
+      "step": 11400
+    },
+    {
+      "epoch": 20.47,
+      "learning_rate": 3.279473369239976e-05,
+      "loss": 0.1364,
+      "step": 11410
+    },
+    {
+      "epoch": 20.48,
+      "learning_rate": 3.2774785557550366e-05,
+      "loss": 0.1183,
+      "step": 11420
+    },
+    {
+      "epoch": 20.5,
+      "learning_rate": 3.275483742270098e-05,
+      "loss": 0.1112,
+      "step": 11430
+    },
+    {
+      "epoch": 20.52,
+      "learning_rate": 3.273488928785159e-05,
+      "loss": 0.1126,
+      "step": 11440
+    },
+    {
+      "epoch": 20.54,
+      "learning_rate": 3.271494115300219e-05,
+      "loss": 0.0929,
+      "step": 11450
+    },
+    {
+      "epoch": 20.56,
+      "learning_rate": 3.26949930181528e-05,
+      "loss": 0.1057,
+      "step": 11460
+    },
+    {
+      "epoch": 20.57,
+      "learning_rate": 3.2675044883303415e-05,
+      "loss": 0.1167,
+      "step": 11470
+    },
+    {
+      "epoch": 20.59,
+      "learning_rate": 3.265509674845402e-05,
+      "loss": 0.1217,
+      "step": 11480
+    },
+    {
+      "epoch": 20.61,
+      "learning_rate": 3.2635148613604626e-05,
+      "loss": 0.1101,
+      "step": 11490
+    },
+    {
+      "epoch": 20.63,
+      "learning_rate": 3.261520047875524e-05,
+      "loss": 0.0885,
+      "step": 11500
+    },
+    {
+      "epoch": 20.65,
+      "learning_rate": 3.2595252343905844e-05,
+      "loss": 0.1036,
+      "step": 11510
+    },
+    {
+      "epoch": 20.66,
+      "learning_rate": 3.2575304209056456e-05,
+      "loss": 0.1268,
+      "step": 11520
+    },
+    {
+      "epoch": 20.68,
+      "learning_rate": 3.255535607420706e-05,
+      "loss": 0.1195,
+      "step": 11530
+    },
+    {
+      "epoch": 20.7,
+      "learning_rate": 3.253540793935767e-05,
+      "loss": 0.1039,
+      "step": 11540
+    },
+    {
+      "epoch": 20.72,
+      "learning_rate": 3.251545980450828e-05,
+      "loss": 0.0855,
+      "step": 11550
+    },
+    {
+      "epoch": 20.74,
+      "learning_rate": 3.249551166965889e-05,
+      "loss": 0.1096,
+      "step": 11560
+    },
+    {
+      "epoch": 20.75,
+      "learning_rate": 3.247556353480949e-05,
+      "loss": 0.1106,
+      "step": 11570
+    },
+    {
+      "epoch": 20.77,
+      "learning_rate": 3.2455615399960104e-05,
+      "loss": 0.1085,
+      "step": 11580
+    },
+    {
+      "epoch": 20.79,
+      "learning_rate": 3.2435667265110716e-05,
+      "loss": 0.126,
+      "step": 11590
+    },
+    {
+      "epoch": 20.81,
+      "learning_rate": 3.241571913026132e-05,
+      "loss": 0.0959,
+      "step": 11600
+    },
+    {
+      "epoch": 20.83,
+      "learning_rate": 3.239577099541193e-05,
+      "loss": 0.1152,
+      "step": 11610
+    },
+    {
+      "epoch": 20.84,
+      "learning_rate": 3.237582286056254e-05,
+      "loss": 0.1252,
+      "step": 11620
+    },
+    {
+      "epoch": 20.86,
+      "learning_rate": 3.2355874725713145e-05,
+      "loss": 0.1151,
+      "step": 11630
+    },
+    {
+      "epoch": 20.88,
+      "learning_rate": 3.233592659086376e-05,
+      "loss": 0.1083,
+      "step": 11640
+    },
+    {
+      "epoch": 20.9,
+      "learning_rate": 3.231597845601436e-05,
+      "loss": 0.1292,
+      "step": 11650
+    },
+    {
+      "epoch": 20.91,
+      "learning_rate": 3.229603032116497e-05,
+      "loss": 0.1201,
+      "step": 11660
+    },
+    {
+      "epoch": 20.93,
+      "learning_rate": 3.227608218631558e-05,
+      "loss": 0.1006,
+      "step": 11670
+    },
+    {
+      "epoch": 20.95,
+      "learning_rate": 3.2256134051466194e-05,
+      "loss": 0.1131,
+      "step": 11680
+    },
+    {
+      "epoch": 20.97,
+      "learning_rate": 3.223618591661679e-05,
+      "loss": 0.1206,
+      "step": 11690
+    },
+    {
+      "epoch": 20.99,
+      "learning_rate": 3.2216237781767405e-05,
+      "loss": 0.0912,
+      "step": 11700
+    },
+    {
+      "epoch": 21.0,
+      "eval_accuracy": {
+        "accuracy": 0.9757793227180982
+      },
+      "eval_f1": {
+        "f1": 0.9743545836218206
+      },
+      "eval_loss": 0.07501205056905746,
+      "eval_precision": {
+        "precision": 0.9744013683930541
+      },
+      "eval_recall": {
+        "recall": 0.9743462910019547
+      },
+      "eval_runtime": 96.9661,
+      "eval_samples_per_second": 183.941,
+      "eval_steps_per_second": 5.755,
+      "step": 11707
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 27850,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 50,
+  "save_steps": 500,
+  "total_flos": 3.723983236417167e+19,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}