diff --git "a/tmp-checkpoint-26754/trainer_state.json" "b/tmp-checkpoint-26754/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/tmp-checkpoint-26754/trainer_state.json"
@@ -0,0 +1,16551 @@
+{
+  "best_metric": 0.05940975993871689,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-crack-detectorVITmain50epochs\\checkpoint-23409",
+  "epoch": 24.0,
+  "eval_steps": 500,
+  "global_step": 26754,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.976660682226213e-08,
+      "loss": 1.3961,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.7953321364452426e-07,
+      "loss": 1.395,
+      "step": 20
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.692998204667864e-07,
+      "loss": 1.3978,
+      "step": 30
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.590664272890485e-07,
+      "loss": 1.3963,
+      "step": 40
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.4883303411131064e-07,
+      "loss": 1.3907,
+      "step": 50
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.385996409335728e-07,
+      "loss": 1.396,
+      "step": 60
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 6.283662477558349e-07,
+      "loss": 1.3885,
+      "step": 70
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 7.18132854578097e-07,
+      "loss": 1.387,
+      "step": 80
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 8.078994614003591e-07,
+      "loss": 1.3894,
+      "step": 90
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 8.976660682226213e-07,
+      "loss": 1.3806,
+      "step": 100
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.874326750448833e-07,
+      "loss": 1.383,
+      "step": 110
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.0771992818671456e-06,
+      "loss": 1.3764,
+      "step": 120
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.1669658886894075e-06,
+      "loss": 1.3688,
+      "step": 130
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.2567324955116697e-06,
+      "loss": 1.3708,
+      "step": 140
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.3464991023339318e-06,
+      "loss": 1.366,
+      "step": 150
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.436265709156194e-06,
+      "loss": 1.371,
+      "step": 160
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.5260323159784561e-06,
+      "loss": 1.3558,
+      "step": 170
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.6157989228007182e-06,
+      "loss": 1.3584,
+      "step": 180
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.7055655296229805e-06,
+      "loss": 1.3409,
+      "step": 190
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.7953321364452425e-06,
+      "loss": 1.346,
+      "step": 200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8850987432675046e-06,
+      "loss": 1.3355,
+      "step": 210
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9748653500897667e-06,
+      "loss": 1.3323,
+      "step": 220
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.064631956912029e-06,
+      "loss": 1.3231,
+      "step": 230
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.1543985637342912e-06,
+      "loss": 1.3129,
+      "step": 240
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.244165170556553e-06,
+      "loss": 1.3178,
+      "step": 250
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.333931777378815e-06,
+      "loss": 1.2953,
+      "step": 260
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.423698384201077e-06,
+      "loss": 1.2938,
+      "step": 270
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.5134649910233395e-06,
+      "loss": 1.2849,
+      "step": 280
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.6032315978456015e-06,
+      "loss": 1.2813,
+      "step": 290
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.6929982046678636e-06,
+      "loss": 1.2811,
+      "step": 300
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.7827648114901257e-06,
+      "loss": 1.2526,
+      "step": 310
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.872531418312388e-06,
+      "loss": 1.2438,
+      "step": 320
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.9622980251346502e-06,
+      "loss": 1.2145,
+      "step": 330
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.0520646319569123e-06,
+      "loss": 1.2146,
+      "step": 340
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.1418312387791743e-06,
+      "loss": 1.1883,
+      "step": 350
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.2315978456014364e-06,
+      "loss": 1.1891,
+      "step": 360
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.321364452423698e-06,
+      "loss": 1.1723,
+      "step": 370
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 3.411131059245961e-06,
+      "loss": 1.1417,
+      "step": 380
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 3.500897666068223e-06,
+      "loss": 1.1338,
+      "step": 390
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 3.590664272890485e-06,
+      "loss": 1.1004,
+      "step": 400
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 3.680430879712747e-06,
+      "loss": 1.0873,
+      "step": 410
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 3.770197486535009e-06,
+      "loss": 1.0358,
+      "step": 420
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 3.859964093357271e-06,
+      "loss": 1.0517,
+      "step": 430
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 3.949730700179533e-06,
+      "loss": 1.0139,
+      "step": 440
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.039497307001795e-06,
+      "loss": 1.0294,
+      "step": 450
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.129263913824058e-06,
+      "loss": 1.0171,
+      "step": 460
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.21903052064632e-06,
+      "loss": 0.9827,
+      "step": 470
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.3087971274685824e-06,
+      "loss": 0.9382,
+      "step": 480
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.3985637342908445e-06,
+      "loss": 0.9145,
+      "step": 490
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.488330341113106e-06,
+      "loss": 0.9276,
+      "step": 500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.578096947935368e-06,
+      "loss": 0.8801,
+      "step": 510
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.66786355475763e-06,
+      "loss": 0.8906,
+      "step": 520
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.757630161579892e-06,
+      "loss": 0.9025,
+      "step": 530
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.847396768402154e-06,
+      "loss": 0.8621,
+      "step": 540
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.937163375224417e-06,
+      "loss": 0.8653,
+      "step": 550
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.026929982046679e-06,
+      "loss": 0.8292,
+      "step": 560
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.116696588868941e-06,
+      "loss": 0.8195,
+      "step": 570
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.206463195691203e-06,
+      "loss": 0.8026,
+      "step": 580
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.296229802513465e-06,
+      "loss": 0.841,
+      "step": 590
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.385996409335727e-06,
+      "loss": 0.778,
+      "step": 600
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.475763016157989e-06,
+      "loss": 0.7706,
+      "step": 610
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.565529622980251e-06,
+      "loss": 0.733,
+      "step": 620
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.655296229802514e-06,
+      "loss": 0.7437,
+      "step": 630
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.745062836624776e-06,
+      "loss": 0.7644,
+      "step": 640
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.834829443447038e-06,
+      "loss": 0.7125,
+      "step": 650
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.9245960502693004e-06,
+      "loss": 0.7146,
+      "step": 660
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 6.0143626570915625e-06,
+      "loss": 0.7201,
+      "step": 670
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.1041292639138246e-06,
+      "loss": 0.6851,
+      "step": 680
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.193895870736087e-06,
+      "loss": 0.6417,
+      "step": 690
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.283662477558349e-06,
+      "loss": 0.6656,
+      "step": 700
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.373429084380611e-06,
+      "loss": 0.6122,
+      "step": 710
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.463195691202873e-06,
+      "loss": 0.6536,
+      "step": 720
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 6.552962298025135e-06,
+      "loss": 0.6315,
+      "step": 730
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 6.642728904847396e-06,
+      "loss": 0.6618,
+      "step": 740
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 6.732495511669659e-06,
+      "loss": 0.6346,
+      "step": 750
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 6.822262118491922e-06,
+      "loss": 0.6866,
+      "step": 760
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.912028725314183e-06,
+      "loss": 0.6093,
+      "step": 770
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 7.001795332136446e-06,
+      "loss": 0.6116,
+      "step": 780
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 7.091561938958707e-06,
+      "loss": 0.602,
+      "step": 790
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 7.18132854578097e-06,
+      "loss": 0.6266,
+      "step": 800
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 7.271095152603231e-06,
+      "loss": 0.6082,
+      "step": 810
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 7.360861759425494e-06,
+      "loss": 0.6194,
+      "step": 820
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 7.4506283662477555e-06,
+      "loss": 0.6277,
+      "step": 830
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 7.540394973070018e-06,
+      "loss": 0.609,
+      "step": 840
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 7.630161579892281e-06,
+      "loss": 0.6252,
+      "step": 850
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 7.719928186714543e-06,
+      "loss": 0.5813,
+      "step": 860
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.809694793536805e-06,
+      "loss": 0.5799,
+      "step": 870
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.899461400359067e-06,
+      "loss": 0.6398,
+      "step": 880
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 7.98922800718133e-06,
+      "loss": 0.5549,
+      "step": 890
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 8.07899461400359e-06,
+      "loss": 0.6023,
+      "step": 900
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 8.168761220825854e-06,
+      "loss": 0.582,
+      "step": 910
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.258527827648117e-06,
+      "loss": 0.544,
+      "step": 920
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.348294434470378e-06,
+      "loss": 0.5912,
+      "step": 930
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.43806104129264e-06,
+      "loss": 0.5461,
+      "step": 940
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.527827648114902e-06,
+      "loss": 0.5238,
+      "step": 950
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 8.617594254937165e-06,
+      "loss": 0.5558,
+      "step": 960
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.707360861759426e-06,
+      "loss": 0.5371,
+      "step": 970
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.797127468581689e-06,
+      "loss": 0.5529,
+      "step": 980
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.88689407540395e-06,
+      "loss": 0.5691,
+      "step": 990
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.976660682226211e-06,
+      "loss": 0.5567,
+      "step": 1000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 9.066427289048474e-06,
+      "loss": 0.5135,
+      "step": 1010
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 9.156193895870736e-06,
+      "loss": 0.5099,
+      "step": 1020
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 9.245960502692998e-06,
+      "loss": 0.5422,
+      "step": 1030
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.33572710951526e-06,
+      "loss": 0.5106,
+      "step": 1040
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 9.425493716337523e-06,
+      "loss": 0.4892,
+      "step": 1050
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 9.515260323159784e-06,
+      "loss": 0.4541,
+      "step": 1060
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.605026929982047e-06,
+      "loss": 0.4675,
+      "step": 1070
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.694793536804308e-06,
+      "loss": 0.5123,
+      "step": 1080
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.784560143626571e-06,
+      "loss": 0.5089,
+      "step": 1090
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.874326750448834e-06,
+      "loss": 0.5233,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.964093357271095e-06,
+      "loss": 0.5622,
+      "step": 1110
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.8736263736263736
+      },
+      "eval_f1": {
+        "f1": 0.8653360894930927
+      },
+      "eval_loss": 0.4108576774597168,
+      "eval_precision": {
+        "precision": 0.8669392826942903
+      },
+      "eval_recall": {
+        "recall": 0.8641606609285752
+      },
+      "eval_runtime": 236.9507,
+      "eval_samples_per_second": 75.273,
+      "eval_steps_per_second": 4.706,
+      "step": 1114
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0053859964093358e-05,
+      "loss": 0.4957,
+      "step": 1120
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0143626570915619e-05,
+      "loss": 0.5729,
+      "step": 1130
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0233393177737882e-05,
+      "loss": 0.4884,
+      "step": 1140
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0323159784560143e-05,
+      "loss": 0.4744,
+      "step": 1150
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.0412926391382406e-05,
+      "loss": 0.4683,
+      "step": 1160
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.0502692998204669e-05,
+      "loss": 0.5,
+      "step": 1170
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.059245960502693e-05,
+      "loss": 0.4649,
+      "step": 1180
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.0682226211849193e-05,
+      "loss": 0.508,
+      "step": 1190
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.0771992818671454e-05,
+      "loss": 0.4898,
+      "step": 1200
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.0861759425493717e-05,
+      "loss": 0.4722,
+      "step": 1210
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.0951526032315979e-05,
+      "loss": 0.4814,
+      "step": 1220
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.1041292639138241e-05,
+      "loss": 0.4505,
+      "step": 1230
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.1131059245960503e-05,
+      "loss": 0.4505,
+      "step": 1240
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.1220825852782766e-05,
+      "loss": 0.4187,
+      "step": 1250
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.1310592459605028e-05,
+      "loss": 0.4603,
+      "step": 1260
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.140035906642729e-05,
+      "loss": 0.4721,
+      "step": 1270
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.1490125673249553e-05,
+      "loss": 0.4368,
+      "step": 1280
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.1579892280071814e-05,
+      "loss": 0.4722,
+      "step": 1290
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.1669658886894077e-05,
+      "loss": 0.4718,
+      "step": 1300
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.1759425493716338e-05,
+      "loss": 0.4899,
+      "step": 1310
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.1849192100538601e-05,
+      "loss": 0.4758,
+      "step": 1320
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.1938958707360862e-05,
+      "loss": 0.4523,
+      "step": 1330
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.2028725314183125e-05,
+      "loss": 0.485,
+      "step": 1340
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.2118491921005388e-05,
+      "loss": 0.4977,
+      "step": 1350
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.2208258527827649e-05,
+      "loss": 0.4305,
+      "step": 1360
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.2298025134649912e-05,
+      "loss": 0.4221,
+      "step": 1370
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.2387791741472173e-05,
+      "loss": 0.4975,
+      "step": 1380
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.2477558348294434e-05,
+      "loss": 0.4506,
+      "step": 1390
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.2567324955116697e-05,
+      "loss": 0.4326,
+      "step": 1400
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.2657091561938959e-05,
+      "loss": 0.4479,
+      "step": 1410
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.2746858168761221e-05,
+      "loss": 0.4468,
+      "step": 1420
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.2836624775583483e-05,
+      "loss": 0.402,
+      "step": 1430
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.2926391382405746e-05,
+      "loss": 0.4711,
+      "step": 1440
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.3016157989228009e-05,
+      "loss": 0.4686,
+      "step": 1450
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.310592459605027e-05,
+      "loss": 0.4639,
+      "step": 1460
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.3195691202872531e-05,
+      "loss": 0.435,
+      "step": 1470
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.3285457809694792e-05,
+      "loss": 0.3868,
+      "step": 1480
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.3375224416517057e-05,
+      "loss": 0.4303,
+      "step": 1490
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.3464991023339318e-05,
+      "loss": 0.432,
+      "step": 1500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.355475763016158e-05,
+      "loss": 0.4206,
+      "step": 1510
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.3644524236983844e-05,
+      "loss": 0.4282,
+      "step": 1520
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.3734290843806105e-05,
+      "loss": 0.4284,
+      "step": 1530
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.3824057450628366e-05,
+      "loss": 0.4096,
+      "step": 1540
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.3913824057450627e-05,
+      "loss": 0.4259,
+      "step": 1550
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.4003590664272892e-05,
+      "loss": 0.4436,
+      "step": 1560
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.4093357271095153e-05,
+      "loss": 0.4453,
+      "step": 1570
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.4183123877917415e-05,
+      "loss": 0.4065,
+      "step": 1580
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.4272890484739679e-05,
+      "loss": 0.4011,
+      "step": 1590
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.436265709156194e-05,
+      "loss": 0.4488,
+      "step": 1600
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.4452423698384202e-05,
+      "loss": 0.4122,
+      "step": 1610
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.4542190305206463e-05,
+      "loss": 0.429,
+      "step": 1620
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.4631956912028727e-05,
+      "loss": 0.4377,
+      "step": 1630
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.4721723518850989e-05,
+      "loss": 0.4186,
+      "step": 1640
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.481149012567325e-05,
+      "loss": 0.4063,
+      "step": 1650
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.4901256732495511e-05,
+      "loss": 0.4055,
+      "step": 1660
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.4991023339317776e-05,
+      "loss": 0.4173,
+      "step": 1670
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.5080789946140037e-05,
+      "loss": 0.4361,
+      "step": 1680
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.5170556552962298e-05,
+      "loss": 0.4528,
+      "step": 1690
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.5260323159784563e-05,
+      "loss": 0.3665,
+      "step": 1700
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.5350089766606824e-05,
+      "loss": 0.3913,
+      "step": 1710
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.5439856373429085e-05,
+      "loss": 0.4002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.5529622980251346e-05,
+      "loss": 0.4004,
+      "step": 1730
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.561938958707361e-05,
+      "loss": 0.4432,
+      "step": 1740
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.5709156193895872e-05,
+      "loss": 0.3743,
+      "step": 1750
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.5798922800718133e-05,
+      "loss": 0.3972,
+      "step": 1760
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.5888689407540398e-05,
+      "loss": 0.368,
+      "step": 1770
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.597845601436266e-05,
+      "loss": 0.4525,
+      "step": 1780
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.606822262118492e-05,
+      "loss": 0.3962,
+      "step": 1790
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.615798922800718e-05,
+      "loss": 0.3888,
+      "step": 1800
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.6247755834829446e-05,
+      "loss": 0.4111,
+      "step": 1810
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.6337522441651707e-05,
+      "loss": 0.4084,
+      "step": 1820
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.642728904847397e-05,
+      "loss": 0.4181,
+      "step": 1830
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.6517055655296233e-05,
+      "loss": 0.4022,
+      "step": 1840
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.6606822262118494e-05,
+      "loss": 0.3886,
+      "step": 1850
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.6696588868940756e-05,
+      "loss": 0.4092,
+      "step": 1860
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.6786355475763017e-05,
+      "loss": 0.3843,
+      "step": 1870
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.687612208258528e-05,
+      "loss": 0.4237,
+      "step": 1880
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.6965888689407543e-05,
+      "loss": 0.3801,
+      "step": 1890
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.7055655296229804e-05,
+      "loss": 0.385,
+      "step": 1900
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.7145421903052065e-05,
+      "loss": 0.4452,
+      "step": 1910
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.723518850987433e-05,
+      "loss": 0.5147,
+      "step": 1920
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.732495511669659e-05,
+      "loss": 0.353,
+      "step": 1930
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.7414721723518852e-05,
+      "loss": 0.3816,
+      "step": 1940
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.7504488330341113e-05,
+      "loss": 0.361,
+      "step": 1950
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.7594254937163378e-05,
+      "loss": 0.3884,
+      "step": 1960
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.768402154398564e-05,
+      "loss": 0.3391,
+      "step": 1970
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.77737881508079e-05,
+      "loss": 0.4211,
+      "step": 1980
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.786355475763016e-05,
+      "loss": 0.3796,
+      "step": 1990
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.7953321364452423e-05,
+      "loss": 0.3881,
+      "step": 2000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8043087971274687e-05,
+      "loss": 0.3904,
+      "step": 2010
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.813285457809695e-05,
+      "loss": 0.3845,
+      "step": 2020
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.822262118491921e-05,
+      "loss": 0.3721,
+      "step": 2030
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.831238779174147e-05,
+      "loss": 0.4141,
+      "step": 2040
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.8402154398563732e-05,
+      "loss": 0.3682,
+      "step": 2050
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.8491921005385997e-05,
+      "loss": 0.3826,
+      "step": 2060
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.8581687612208258e-05,
+      "loss": 0.3588,
+      "step": 2070
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.867145421903052e-05,
+      "loss": 0.3663,
+      "step": 2080
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.8761220825852784e-05,
+      "loss": 0.4186,
+      "step": 2090
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8850987432675045e-05,
+      "loss": 0.3756,
+      "step": 2100
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.8940754039497306e-05,
+      "loss": 0.3655,
+      "step": 2110
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.9030520646319568e-05,
+      "loss": 0.4158,
+      "step": 2120
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.9120287253141832e-05,
+      "loss": 0.3985,
+      "step": 2130
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.9210053859964093e-05,
+      "loss": 0.3955,
+      "step": 2140
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.9299820466786355e-05,
+      "loss": 0.3757,
+      "step": 2150
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.9389587073608616e-05,
+      "loss": 0.3712,
+      "step": 2160
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.947935368043088e-05,
+      "loss": 0.3723,
+      "step": 2170
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.9569120287253142e-05,
+      "loss": 0.367,
+      "step": 2180
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.9658886894075403e-05,
+      "loss": 0.3659,
+      "step": 2190
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.9748653500897668e-05,
+      "loss": 0.3752,
+      "step": 2200
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.983842010771993e-05,
+      "loss": 0.318,
+      "step": 2210
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.992818671454219e-05,
+      "loss": 0.3696,
+      "step": 2220
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.9249271137026239
+      },
+      "eval_f1": {
+        "f1": 0.9207602119769538
+      },
+      "eval_loss": 0.2492757886648178,
+      "eval_precision": {
+        "precision": 0.9199949527238288
+      },
+      "eval_recall": {
+        "recall": 0.9219224853720074
+      },
+      "eval_runtime": 166.1808,
+      "eval_samples_per_second": 107.329,
+      "eval_steps_per_second": 6.71,
+      "step": 2229
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.001795332136445e-05,
+      "loss": 0.3681,
+      "step": 2230
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.0107719928186716e-05,
+      "loss": 0.321,
+      "step": 2240
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.0197486535008977e-05,
+      "loss": 0.332,
+      "step": 2250
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.0287253141831238e-05,
+      "loss": 0.3343,
+      "step": 2260
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.0377019748653503e-05,
+      "loss": 0.3341,
+      "step": 2270
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.0466786355475764e-05,
+      "loss": 0.3473,
+      "step": 2280
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.0556552962298025e-05,
+      "loss": 0.3479,
+      "step": 2290
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.0646319569120286e-05,
+      "loss": 0.3235,
+      "step": 2300
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.073608617594255e-05,
+      "loss": 0.3455,
+      "step": 2310
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.0825852782764812e-05,
+      "loss": 0.3675,
+      "step": 2320
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.0915619389587073e-05,
+      "loss": 0.3267,
+      "step": 2330
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.1005385996409338e-05,
+      "loss": 0.3588,
+      "step": 2340
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.10951526032316e-05,
+      "loss": 0.3762,
+      "step": 2350
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.118491921005386e-05,
+      "loss": 0.3474,
+      "step": 2360
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.1274685816876122e-05,
+      "loss": 0.3614,
+      "step": 2370
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.1364452423698386e-05,
+      "loss": 0.3611,
+      "step": 2380
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.1454219030520648e-05,
+      "loss": 0.4051,
+      "step": 2390
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.154398563734291e-05,
+      "loss": 0.3314,
+      "step": 2400
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.163375224416517e-05,
+      "loss": 0.3427,
+      "step": 2410
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.1723518850987435e-05,
+      "loss": 0.2886,
+      "step": 2420
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.1813285457809696e-05,
+      "loss": 0.3262,
+      "step": 2430
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.1903052064631957e-05,
+      "loss": 0.3516,
+      "step": 2440
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 2.199281867145422e-05,
+      "loss": 0.3744,
+      "step": 2450
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.2082585278276483e-05,
+      "loss": 0.3545,
+      "step": 2460
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.2172351885098744e-05,
+      "loss": 0.2993,
+      "step": 2470
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.2262118491921005e-05,
+      "loss": 0.3412,
+      "step": 2480
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.235188509874327e-05,
+      "loss": 0.3823,
+      "step": 2490
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.244165170556553e-05,
+      "loss": 0.3206,
+      "step": 2500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.2531418312387792e-05,
+      "loss": 0.2969,
+      "step": 2510
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.2621184919210057e-05,
+      "loss": 0.3306,
+      "step": 2520
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.2710951526032318e-05,
+      "loss": 0.3641,
+      "step": 2530
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.280071813285458e-05,
+      "loss": 0.3588,
+      "step": 2540
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.289048473967684e-05,
+      "loss": 0.3632,
+      "step": 2550
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.2980251346499105e-05,
+      "loss": 0.3072,
+      "step": 2560
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.3070017953321366e-05,
+      "loss": 0.3606,
+      "step": 2570
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.3159784560143628e-05,
+      "loss": 0.4114,
+      "step": 2580
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.3249551166965892e-05,
+      "loss": 0.282,
+      "step": 2590
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.3339317773788153e-05,
+      "loss": 0.3336,
+      "step": 2600
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.3429084380610415e-05,
+      "loss": 0.3406,
+      "step": 2610
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.3518850987432676e-05,
+      "loss": 0.3433,
+      "step": 2620
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.360861759425494e-05,
+      "loss": 0.3169,
+      "step": 2630
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.3698384201077202e-05,
+      "loss": 0.3855,
+      "step": 2640
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.3788150807899463e-05,
+      "loss": 0.3036,
+      "step": 2650
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.3877917414721724e-05,
+      "loss": 0.3076,
+      "step": 2660
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.396768402154399e-05,
+      "loss": 0.3052,
+      "step": 2670
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.405745062836625e-05,
+      "loss": 0.3253,
+      "step": 2680
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 2.414721723518851e-05,
+      "loss": 0.3235,
+      "step": 2690
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.4236983842010776e-05,
+      "loss": 0.2835,
+      "step": 2700
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.4326750448833037e-05,
+      "loss": 0.2824,
+      "step": 2710
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.4416517055655298e-05,
+      "loss": 0.3183,
+      "step": 2720
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.450628366247756e-05,
+      "loss": 0.3281,
+      "step": 2730
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.4596050269299824e-05,
+      "loss": 0.2939,
+      "step": 2740
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 2.4685816876122085e-05,
+      "loss": 0.3679,
+      "step": 2750
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.4775583482944346e-05,
+      "loss": 0.3165,
+      "step": 2760
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.4865350089766608e-05,
+      "loss": 0.3146,
+      "step": 2770
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.495511669658887e-05,
+      "loss": 0.2981,
+      "step": 2780
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 2.5044883303411134e-05,
+      "loss": 0.3099,
+      "step": 2790
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.5134649910233395e-05,
+      "loss": 0.267,
+      "step": 2800
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.5224416517055656e-05,
+      "loss": 0.3317,
+      "step": 2810
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.5314183123877917e-05,
+      "loss": 0.3241,
+      "step": 2820
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.5403949730700182e-05,
+      "loss": 0.371,
+      "step": 2830
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.5493716337522443e-05,
+      "loss": 0.3431,
+      "step": 2840
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.5583482944344704e-05,
+      "loss": 0.3218,
+      "step": 2850
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.5673249551166965e-05,
+      "loss": 0.3222,
+      "step": 2860
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.5763016157989227e-05,
+      "loss": 0.3176,
+      "step": 2870
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.585278276481149e-05,
+      "loss": 0.3256,
+      "step": 2880
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.5942549371633752e-05,
+      "loss": 0.2971,
+      "step": 2890
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.6032315978456017e-05,
+      "loss": 0.253,
+      "step": 2900
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.6122082585278278e-05,
+      "loss": 0.3261,
+      "step": 2910
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.621184919210054e-05,
+      "loss": 0.3325,
+      "step": 2920
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.63016157989228e-05,
+      "loss": 0.3005,
+      "step": 2930
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.6391382405745062e-05,
+      "loss": 0.2897,
+      "step": 2940
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.6481149012567323e-05,
+      "loss": 0.3249,
+      "step": 2950
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.6570915619389584e-05,
+      "loss": 0.3647,
+      "step": 2960
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.6660682226211852e-05,
+      "loss": 0.3034,
+      "step": 2970
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.6750448833034114e-05,
+      "loss": 0.2735,
+      "step": 2980
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.6840215439856375e-05,
+      "loss": 0.3122,
+      "step": 2990
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.6929982046678636e-05,
+      "loss": 0.2622,
+      "step": 3000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.7019748653500897e-05,
+      "loss": 0.322,
+      "step": 3010
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.710951526032316e-05,
+      "loss": 0.2671,
+      "step": 3020
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.719928186714542e-05,
+      "loss": 0.3392,
+      "step": 3030
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.7289048473967688e-05,
+      "loss": 0.2507,
+      "step": 3040
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.737881508078995e-05,
+      "loss": 0.294,
+      "step": 3050
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.746858168761221e-05,
+      "loss": 0.3119,
+      "step": 3060
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.755834829443447e-05,
+      "loss": 0.2956,
+      "step": 3070
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.7648114901256732e-05,
+      "loss": 0.2744,
+      "step": 3080
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.7737881508078994e-05,
+      "loss": 0.279,
+      "step": 3090
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.7827648114901255e-05,
+      "loss": 0.379,
+      "step": 3100
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 2.7917414721723523e-05,
+      "loss": 0.2667,
+      "step": 3110
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.8007181328545784e-05,
+      "loss": 0.2732,
+      "step": 3120
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 2.8096947935368045e-05,
+      "loss": 0.311,
+      "step": 3130
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.8186714542190307e-05,
+      "loss": 0.3404,
+      "step": 3140
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.8276481149012568e-05,
+      "loss": 0.2479,
+      "step": 3150
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.836624775583483e-05,
+      "loss": 0.3057,
+      "step": 3160
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 2.845601436265709e-05,
+      "loss": 0.3075,
+      "step": 3170
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.8545780969479358e-05,
+      "loss": 0.3008,
+      "step": 3180
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.863554757630162e-05,
+      "loss": 0.2372,
+      "step": 3190
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.872531418312388e-05,
+      "loss": 0.2617,
+      "step": 3200
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.8815080789946142e-05,
+      "loss": 0.3,
+      "step": 3210
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.8904847396768403e-05,
+      "loss": 0.2749,
+      "step": 3220
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.8994614003590664e-05,
+      "loss": 0.2742,
+      "step": 3230
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.9084380610412926e-05,
+      "loss": 0.2844,
+      "step": 3240
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.9174147217235194e-05,
+      "loss": 0.2909,
+      "step": 3250
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.9263913824057455e-05,
+      "loss": 0.3045,
+      "step": 3260
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 2.9353680430879716e-05,
+      "loss": 0.2875,
+      "step": 3270
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.9443447037701977e-05,
+      "loss": 0.2963,
+      "step": 3280
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.953321364452424e-05,
+      "loss": 0.314,
+      "step": 3290
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.96229802513465e-05,
+      "loss": 0.3005,
+      "step": 3300
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.971274685816876e-05,
+      "loss": 0.3127,
+      "step": 3310
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.9802513464991022e-05,
+      "loss": 0.3113,
+      "step": 3320
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.989228007181329e-05,
+      "loss": 0.2886,
+      "step": 3330
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.998204667863555e-05,
+      "loss": 0.321,
+      "step": 3340
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.9410742318905584
+      },
+      "eval_f1": {
+        "f1": 0.9378351696354423
+      },
+      "eval_loss": 0.18490658700466156,
+      "eval_precision": {
+        "precision": 0.9376663800198108
+      },
+      "eval_recall": {
+        "recall": 0.9381803360170522
+      },
+      "eval_runtime": 165.2584,
+      "eval_samples_per_second": 107.928,
+      "eval_steps_per_second": 6.747,
+      "step": 3344
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.0071813285457812e-05,
+      "loss": 0.3252,
+      "step": 3350
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.0161579892280074e-05,
+      "loss": 0.2945,
+      "step": 3360
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 3.0251346499102335e-05,
+      "loss": 0.2584,
+      "step": 3370
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.0341113105924596e-05,
+      "loss": 0.2878,
+      "step": 3380
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.0430879712746857e-05,
+      "loss": 0.2618,
+      "step": 3390
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.0520646319569125e-05,
+      "loss": 0.3356,
+      "step": 3400
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 3.061041292639138e-05,
+      "loss": 0.2706,
+      "step": 3410
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.070017953321365e-05,
+      "loss": 0.2821,
+      "step": 3420
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.0789946140035906e-05,
+      "loss": 0.2634,
+      "step": 3430
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.087971274685817e-05,
+      "loss": 0.2927,
+      "step": 3440
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.096947935368043e-05,
+      "loss": 0.2406,
+      "step": 3450
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.105924596050269e-05,
+      "loss": 0.2871,
+      "step": 3460
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.114901256732496e-05,
+      "loss": 0.2389,
+      "step": 3470
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.123877917414722e-05,
+      "loss": 0.2782,
+      "step": 3480
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.132854578096948e-05,
+      "loss": 0.2864,
+      "step": 3490
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.1418312387791744e-05,
+      "loss": 0.29,
+      "step": 3500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.1508078994614e-05,
+      "loss": 0.2089,
+      "step": 3510
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.159784560143627e-05,
+      "loss": 0.2716,
+      "step": 3520
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.1687612208258525e-05,
+      "loss": 0.2981,
+      "step": 3530
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.1777378815080796e-05,
+      "loss": 0.2265,
+      "step": 3540
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.1867145421903054e-05,
+      "loss": 0.1997,
+      "step": 3550
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.195691202872532e-05,
+      "loss": 0.2707,
+      "step": 3560
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.2046678635547576e-05,
+      "loss": 0.2917,
+      "step": 3570
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.213644524236984e-05,
+      "loss": 0.2966,
+      "step": 3580
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 3.22262118491921e-05,
+      "loss": 0.265,
+      "step": 3590
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.231597845601436e-05,
+      "loss": 0.2386,
+      "step": 3600
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.240574506283663e-05,
+      "loss": 0.2613,
+      "step": 3610
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.249551166965889e-05,
+      "loss": 0.2737,
+      "step": 3620
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.258527827648115e-05,
+      "loss": 0.2965,
+      "step": 3630
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.2675044883303415e-05,
+      "loss": 0.2683,
+      "step": 3640
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.276481149012567e-05,
+      "loss": 0.2379,
+      "step": 3650
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.285457809694794e-05,
+      "loss": 0.3282,
+      "step": 3660
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.2944344703770195e-05,
+      "loss": 0.2589,
+      "step": 3670
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.3034111310592466e-05,
+      "loss": 0.2955,
+      "step": 3680
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.3123877917414724e-05,
+      "loss": 0.2584,
+      "step": 3690
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.321364452423699e-05,
+      "loss": 0.2643,
+      "step": 3700
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.330341113105925e-05,
+      "loss": 0.2987,
+      "step": 3710
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.339317773788151e-05,
+      "loss": 0.3,
+      "step": 3720
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.348294434470377e-05,
+      "loss": 0.2488,
+      "step": 3730
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.3572710951526034e-05,
+      "loss": 0.252,
+      "step": 3740
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.36624775583483e-05,
+      "loss": 0.286,
+      "step": 3750
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.375224416517056e-05,
+      "loss": 0.2801,
+      "step": 3760
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.384201077199282e-05,
+      "loss": 0.2463,
+      "step": 3770
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.3931777378815085e-05,
+      "loss": 0.2824,
+      "step": 3780
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.402154398563734e-05,
+      "loss": 0.2847,
+      "step": 3790
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.411131059245961e-05,
+      "loss": 0.3061,
+      "step": 3800
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.4201077199281866e-05,
+      "loss": 0.2799,
+      "step": 3810
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.429084380610413e-05,
+      "loss": 0.2532,
+      "step": 3820
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.4380610412926395e-05,
+      "loss": 0.2761,
+      "step": 3830
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.447037701974866e-05,
+      "loss": 0.3087,
+      "step": 3840
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 3.456014362657092e-05,
+      "loss": 0.2841,
+      "step": 3850
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.464991023339318e-05,
+      "loss": 0.2191,
+      "step": 3860
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.473967684021544e-05,
+      "loss": 0.328,
+      "step": 3870
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.4829443447037704e-05,
+      "loss": 0.2377,
+      "step": 3880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 3.491921005385996e-05,
+      "loss": 0.2844,
+      "step": 3890
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.500897666068223e-05,
+      "loss": 0.2944,
+      "step": 3900
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.509874326750449e-05,
+      "loss": 0.2371,
+      "step": 3910
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.5188509874326756e-05,
+      "loss": 0.2552,
+      "step": 3920
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.5278276481149014e-05,
+      "loss": 0.2492,
+      "step": 3930
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.536804308797128e-05,
+      "loss": 0.253,
+      "step": 3940
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 3.5457809694793536e-05,
+      "loss": 0.2794,
+      "step": 3950
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.55475763016158e-05,
+      "loss": 0.2253,
+      "step": 3960
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 3.5637342908438065e-05,
+      "loss": 0.2894,
+      "step": 3970
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.572710951526032e-05,
+      "loss": 0.2076,
+      "step": 3980
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.581687612208259e-05,
+      "loss": 0.2241,
+      "step": 3990
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.5906642728904846e-05,
+      "loss": 0.2547,
+      "step": 4000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.599640933572711e-05,
+      "loss": 0.2471,
+      "step": 4010
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.6086175942549375e-05,
+      "loss": 0.2585,
+      "step": 4020
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.617594254937163e-05,
+      "loss": 0.2617,
+      "step": 4030
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.62657091561939e-05,
+      "loss": 0.2724,
+      "step": 4040
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.635547576301616e-05,
+      "loss": 0.2353,
+      "step": 4050
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.644524236983842e-05,
+      "loss": 0.2288,
+      "step": 4060
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.6535008976660684e-05,
+      "loss": 0.2431,
+      "step": 4070
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.662477558348294e-05,
+      "loss": 0.2693,
+      "step": 4080
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.671454219030521e-05,
+      "loss": 0.2903,
+      "step": 4090
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.6804308797127465e-05,
+      "loss": 0.2658,
+      "step": 4100
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.6894075403949736e-05,
+      "loss": 0.2175,
+      "step": 4110
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.6983842010771994e-05,
+      "loss": 0.3072,
+      "step": 4120
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.707360861759426e-05,
+      "loss": 0.305,
+      "step": 4130
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.7163375224416516e-05,
+      "loss": 0.2475,
+      "step": 4140
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.725314183123878e-05,
+      "loss": 0.2682,
+      "step": 4150
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.734290843806104e-05,
+      "loss": 0.2593,
+      "step": 4160
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.74326750448833e-05,
+      "loss": 0.3293,
+      "step": 4170
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.752244165170557e-05,
+      "loss": 0.2422,
+      "step": 4180
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.761220825852783e-05,
+      "loss": 0.2018,
+      "step": 4190
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.770197486535009e-05,
+      "loss": 0.2802,
+      "step": 4200
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.7791741472172355e-05,
+      "loss": 0.2298,
+      "step": 4210
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.788150807899461e-05,
+      "loss": 0.2512,
+      "step": 4220
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.797127468581688e-05,
+      "loss": 0.2503,
+      "step": 4230
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.8061041292639135e-05,
+      "loss": 0.2244,
+      "step": 4240
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.815080789946141e-05,
+      "loss": 0.2456,
+      "step": 4250
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.8240574506283664e-05,
+      "loss": 0.2085,
+      "step": 4260
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.833034111310593e-05,
+      "loss": 0.2367,
+      "step": 4270
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.842010771992819e-05,
+      "loss": 0.2174,
+      "step": 4280
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.850987432675045e-05,
+      "loss": 0.2256,
+      "step": 4290
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.859964093357271e-05,
+      "loss": 0.2344,
+      "step": 4300
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.8689407540394974e-05,
+      "loss": 0.2107,
+      "step": 4310
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.877917414721723e-05,
+      "loss": 0.2797,
+      "step": 4320
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.88689407540395e-05,
+      "loss": 0.2568,
+      "step": 4330
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.895870736086176e-05,
+      "loss": 0.2312,
+      "step": 4340
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.9048473967684026e-05,
+      "loss": 0.2269,
+      "step": 4350
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.9138240574506283e-05,
+      "loss": 0.2412,
+      "step": 4360
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 3.922800718132855e-05,
+      "loss": 0.2626,
+      "step": 4370
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.9317773788150806e-05,
+      "loss": 0.2995,
+      "step": 4380
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.940754039497307e-05,
+      "loss": 0.2435,
+      "step": 4390
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.9497307001795335e-05,
+      "loss": 0.2317,
+      "step": 4400
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.95870736086176e-05,
+      "loss": 0.1932,
+      "step": 4410
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.967684021543986e-05,
+      "loss": 0.2498,
+      "step": 4420
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.976660682226212e-05,
+      "loss": 0.2792,
+      "step": 4430
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.985637342908438e-05,
+      "loss": 0.2554,
+      "step": 4440
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 3.9946140035906645e-05,
+      "loss": 0.3403,
+      "step": 4450
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.9382148463781117
+      },
+      "eval_f1": {
+        "f1": 0.9358183038090213
+      },
+      "eval_loss": 0.18279214203357697,
+      "eval_precision": {
+        "precision": 0.9351989305945833
+      },
+      "eval_recall": {
+        "recall": 0.9374254756058774
+      },
+      "eval_runtime": 165.3018,
+      "eval_samples_per_second": 107.9,
+      "eval_steps_per_second": 6.745,
+      "step": 4459
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.00359066427289e-05,
+      "loss": 0.2936,
+      "step": 4460
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 4.0125673249551174e-05,
+      "loss": 0.2723,
+      "step": 4470
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 4.021543985637343e-05,
+      "loss": 0.2307,
+      "step": 4480
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 4.0305206463195696e-05,
+      "loss": 0.2567,
+      "step": 4490
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.0394973070017954e-05,
+      "loss": 0.2273,
+      "step": 4500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.048473967684022e-05,
+      "loss": 0.2638,
+      "step": 4510
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 4.0574506283662476e-05,
+      "loss": 0.2625,
+      "step": 4520
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 4.066427289048474e-05,
+      "loss": 0.2432,
+      "step": 4530
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 4.0754039497307006e-05,
+      "loss": 0.2345,
+      "step": 4540
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.084380610412927e-05,
+      "loss": 0.2272,
+      "step": 4550
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 4.093357271095153e-05,
+      "loss": 0.2217,
+      "step": 4560
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 4.102333931777379e-05,
+      "loss": 0.2195,
+      "step": 4570
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 4.111310592459605e-05,
+      "loss": 0.2637,
+      "step": 4580
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.1202872531418315e-05,
+      "loss": 0.2668,
+      "step": 4590
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 4.129263913824057e-05,
+      "loss": 0.2742,
+      "step": 4600
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.1382405745062844e-05,
+      "loss": 0.2128,
+      "step": 4610
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 4.14721723518851e-05,
+      "loss": 0.241,
+      "step": 4620
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 4.156193895870737e-05,
+      "loss": 0.2332,
+      "step": 4630
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.1651705565529625e-05,
+      "loss": 0.2359,
+      "step": 4640
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 4.174147217235189e-05,
+      "loss": 0.2219,
+      "step": 4650
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 4.183123877917415e-05,
+      "loss": 0.2325,
+      "step": 4660
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 4.192100538599641e-05,
+      "loss": 0.2627,
+      "step": 4670
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.2010771992818676e-05,
+      "loss": 0.196,
+      "step": 4680
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 4.2100538599640934e-05,
+      "loss": 0.234,
+      "step": 4690
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 4.21903052064632e-05,
+      "loss": 0.1998,
+      "step": 4700
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.228007181328546e-05,
+      "loss": 0.2579,
+      "step": 4710
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 4.236983842010772e-05,
+      "loss": 0.2678,
+      "step": 4720
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.2459605026929986e-05,
+      "loss": 0.2431,
+      "step": 4730
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 4.2549371633752244e-05,
+      "loss": 0.2274,
+      "step": 4740
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 4.263913824057451e-05,
+      "loss": 0.2325,
+      "step": 4750
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.272890484739677e-05,
+      "loss": 0.2121,
+      "step": 4760
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 4.281867145421903e-05,
+      "loss": 0.2466,
+      "step": 4770
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 4.2908438061041295e-05,
+      "loss": 0.2331,
+      "step": 4780
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 4.299820466786356e-05,
+      "loss": 0.221,
+      "step": 4790
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.308797127468582e-05,
+      "loss": 0.2563,
+      "step": 4800
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.317773788150808e-05,
+      "loss": 0.2564,
+      "step": 4810
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 4.326750448833034e-05,
+      "loss": 0.2369,
+      "step": 4820
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 4.3357271095152605e-05,
+      "loss": 0.1929,
+      "step": 4830
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 4.344703770197487e-05,
+      "loss": 0.2639,
+      "step": 4840
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.353680430879713e-05,
+      "loss": 0.232,
+      "step": 4850
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 4.362657091561939e-05,
+      "loss": 0.2453,
+      "step": 4860
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 4.371633752244165e-05,
+      "loss": 0.2145,
+      "step": 4870
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 4.3806104129263914e-05,
+      "loss": 0.2249,
+      "step": 4880
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.389587073608618e-05,
+      "loss": 0.2861,
+      "step": 4890
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.398563734290844e-05,
+      "loss": 0.2452,
+      "step": 4900
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 4.40754039497307e-05,
+      "loss": 0.2512,
+      "step": 4910
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 4.4165170556552966e-05,
+      "loss": 0.2373,
+      "step": 4920
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 4.4254937163375224e-05,
+      "loss": 0.2262,
+      "step": 4930
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.434470377019749e-05,
+      "loss": 0.2243,
+      "step": 4940
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 4.4434470377019746e-05,
+      "loss": 0.2228,
+      "step": 4950
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 4.452423698384201e-05,
+      "loss": 0.2331,
+      "step": 4960
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 4.4614003590664275e-05,
+      "loss": 0.2706,
+      "step": 4970
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.470377019748654e-05,
+      "loss": 0.2326,
+      "step": 4980
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 4.47935368043088e-05,
+      "loss": 0.3095,
+      "step": 4990
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.488330341113106e-05,
+      "loss": 0.2106,
+      "step": 5000
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 4.497307001795332e-05,
+      "loss": 0.2494,
+      "step": 5010
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 4.5062836624775585e-05,
+      "loss": 0.234,
+      "step": 5020
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.515260323159784e-05,
+      "loss": 0.283,
+      "step": 5030
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.5242369838420114e-05,
+      "loss": 0.2626,
+      "step": 5040
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.533213644524237e-05,
+      "loss": 0.2352,
+      "step": 5050
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.5421903052064636e-05,
+      "loss": 0.2748,
+      "step": 5060
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.5511669658886894e-05,
+      "loss": 0.2849,
+      "step": 5070
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.560143626570916e-05,
+      "loss": 0.2296,
+      "step": 5080
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.5691202872531417e-05,
+      "loss": 0.2231,
+      "step": 5090
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.578096947935368e-05,
+      "loss": 0.2526,
+      "step": 5100
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.5870736086175946e-05,
+      "loss": 0.2441,
+      "step": 5110
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.596050269299821e-05,
+      "loss": 0.236,
+      "step": 5120
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 4.605026929982047e-05,
+      "loss": 0.2485,
+      "step": 5130
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 4.614003590664273e-05,
+      "loss": 0.1969,
+      "step": 5140
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 4.622980251346499e-05,
+      "loss": 0.2454,
+      "step": 5150
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6319569120287255e-05,
+      "loss": 0.2486,
+      "step": 5160
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 4.640933572710951e-05,
+      "loss": 0.2304,
+      "step": 5170
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 4.6499102333931784e-05,
+      "loss": 0.2321,
+      "step": 5180
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.658886894075404e-05,
+      "loss": 0.2205,
+      "step": 5190
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 4.667863554757631e-05,
+      "loss": 0.2107,
+      "step": 5200
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.6768402154398565e-05,
+      "loss": 0.2215,
+      "step": 5210
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 4.685816876122083e-05,
+      "loss": 0.222,
+      "step": 5220
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 4.694793536804309e-05,
+      "loss": 0.2271,
+      "step": 5230
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 4.703770197486535e-05,
+      "loss": 0.2342,
+      "step": 5240
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.7127468581687616e-05,
+      "loss": 0.2478,
+      "step": 5250
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 4.721723518850988e-05,
+      "loss": 0.2691,
+      "step": 5260
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 4.730700179533214e-05,
+      "loss": 0.2059,
+      "step": 5270
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 4.7396768402154403e-05,
+      "loss": 0.2085,
+      "step": 5280
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.748653500897666e-05,
+      "loss": 0.1888,
+      "step": 5290
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.7576301615798926e-05,
+      "loss": 0.2933,
+      "step": 5300
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 4.7666068222621184e-05,
+      "loss": 0.2519,
+      "step": 5310
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 4.775583482944345e-05,
+      "loss": 0.2142,
+      "step": 5320
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.784560143626571e-05,
+      "loss": 0.1972,
+      "step": 5330
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 4.793536804308798e-05,
+      "loss": 0.2134,
+      "step": 5340
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 4.8025134649910235e-05,
+      "loss": 0.2924,
+      "step": 5350
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 4.81149012567325e-05,
+      "loss": 0.2207,
+      "step": 5360
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.820466786355476e-05,
+      "loss": 0.2369,
+      "step": 5370
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 4.829443447037702e-05,
+      "loss": 0.242,
+      "step": 5380
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.838420107719928e-05,
+      "loss": 0.2023,
+      "step": 5390
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 4.847396768402155e-05,
+      "loss": 0.2413,
+      "step": 5400
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 4.856373429084381e-05,
+      "loss": 0.2757,
+      "step": 5410
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.8653500897666074e-05,
+      "loss": 0.2481,
+      "step": 5420
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.874326750448833e-05,
+      "loss": 0.2487,
+      "step": 5430
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 4.8833034111310596e-05,
+      "loss": 0.1747,
+      "step": 5440
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.8922800718132854e-05,
+      "loss": 0.2537,
+      "step": 5450
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.901256732495512e-05,
+      "loss": 0.2722,
+      "step": 5460
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 4.9102333931777383e-05,
+      "loss": 0.2225,
+      "step": 5470
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.919210053859965e-05,
+      "loss": 0.1949,
+      "step": 5480
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 4.9281867145421906e-05,
+      "loss": 0.2416,
+      "step": 5490
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 4.937163375224417e-05,
+      "loss": 0.2197,
+      "step": 5500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.946140035906643e-05,
+      "loss": 0.2939,
+      "step": 5510
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 4.955116696588869e-05,
+      "loss": 0.2786,
+      "step": 5520
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 4.964093357271095e-05,
+      "loss": 0.2187,
+      "step": 5530
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.9730700179533215e-05,
+      "loss": 0.2518,
+      "step": 5540
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.982046678635548e-05,
+      "loss": 0.205,
+      "step": 5550
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 4.991023339317774e-05,
+      "loss": 0.2277,
+      "step": 5560
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 5e-05,
+      "loss": 0.2399,
+      "step": 5570
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.9539694998878673
+      },
+      "eval_f1": {
+        "f1": 0.9516195088904471
+      },
+      "eval_loss": 0.1283472627401352,
+      "eval_precision": {
+        "precision": 0.9525864748624857
+      },
+      "eval_recall": {
+        "recall": 0.9506985041313745
+      },
+      "eval_runtime": 164.7789,
+      "eval_samples_per_second": 108.242,
+      "eval_steps_per_second": 6.767,
+      "step": 5573
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.999002593257531e-05,
+      "loss": 0.2288,
+      "step": 5580
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 4.9980051865150615e-05,
+      "loss": 0.2216,
+      "step": 5590
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.9970077797725914e-05,
+      "loss": 0.3012,
+      "step": 5600
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 4.9960103730301214e-05,
+      "loss": 0.1937,
+      "step": 5610
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 4.995012966287652e-05,
+      "loss": 0.1837,
+      "step": 5620
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 4.9940155595451826e-05,
+      "loss": 0.2226,
+      "step": 5630
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.993018152802713e-05,
+      "loss": 0.2332,
+      "step": 5640
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 4.992020746060244e-05,
+      "loss": 0.2046,
+      "step": 5650
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 4.991023339317774e-05,
+      "loss": 0.219,
+      "step": 5660
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 4.9900259325753044e-05,
+      "loss": 0.2049,
+      "step": 5670
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.989028525832835e-05,
+      "loss": 0.2299,
+      "step": 5680
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.988031119090365e-05,
+      "loss": 0.2195,
+      "step": 5690
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 4.9870337123478956e-05,
+      "loss": 0.221,
+      "step": 5700
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 4.986036305605426e-05,
+      "loss": 0.2159,
+      "step": 5710
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 4.985038898862957e-05,
+      "loss": 0.2517,
+      "step": 5720
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.984041492120487e-05,
+      "loss": 0.1899,
+      "step": 5730
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 4.9830440853780174e-05,
+      "loss": 0.188,
+      "step": 5740
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 4.982046678635548e-05,
+      "loss": 0.2629,
+      "step": 5750
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 4.9810492718930786e-05,
+      "loss": 0.1955,
+      "step": 5760
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.9800518651506086e-05,
+      "loss": 0.211,
+      "step": 5770
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.979054458408139e-05,
+      "loss": 0.2094,
+      "step": 5780
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 4.978057051665669e-05,
+      "loss": 0.1961,
+      "step": 5790
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 4.9770596449232e-05,
+      "loss": 0.2094,
+      "step": 5800
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 4.9760622381807304e-05,
+      "loss": 0.1873,
+      "step": 5810
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.975064831438261e-05,
+      "loss": 0.2139,
+      "step": 5820
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 4.9740674246957916e-05,
+      "loss": 0.2314,
+      "step": 5830
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 4.9730700179533215e-05,
+      "loss": 0.2177,
+      "step": 5840
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.9720726112108515e-05,
+      "loss": 0.2068,
+      "step": 5850
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 4.971075204468382e-05,
+      "loss": 0.1825,
+      "step": 5860
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.970077797725913e-05,
+      "loss": 0.2621,
+      "step": 5870
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 4.969080390983443e-05,
+      "loss": 0.1762,
+      "step": 5880
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 4.968082984240974e-05,
+      "loss": 0.2026,
+      "step": 5890
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.967085577498504e-05,
+      "loss": 0.1896,
+      "step": 5900
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 4.9660881707560345e-05,
+      "loss": 0.2083,
+      "step": 5910
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 4.965090764013565e-05,
+      "loss": 0.2006,
+      "step": 5920
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 4.964093357271095e-05,
+      "loss": 0.2536,
+      "step": 5930
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.963095950528626e-05,
+      "loss": 0.233,
+      "step": 5940
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 4.962098543786156e-05,
+      "loss": 0.1991,
+      "step": 5950
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 4.961101137043687e-05,
+      "loss": 0.1913,
+      "step": 5960
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.960103730301217e-05,
+      "loss": 0.2553,
+      "step": 5970
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 4.9591063235587475e-05,
+      "loss": 0.22,
+      "step": 5980
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.958108916816278e-05,
+      "loss": 0.1851,
+      "step": 5990
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 4.957111510073809e-05,
+      "loss": 0.2366,
+      "step": 6000
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 4.956114103331339e-05,
+      "loss": 0.1834,
+      "step": 6010
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 4.955116696588869e-05,
+      "loss": 0.1908,
+      "step": 6020
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.954119289846399e-05,
+      "loss": 0.1859,
+      "step": 6030
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 4.95312188310393e-05,
+      "loss": 0.2187,
+      "step": 6040
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 4.9521244763614605e-05,
+      "loss": 0.1675,
+      "step": 6050
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 4.951127069618991e-05,
+      "loss": 0.227,
+      "step": 6060
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.950129662876522e-05,
+      "loss": 0.2286,
+      "step": 6070
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.9491322561340517e-05,
+      "loss": 0.2063,
+      "step": 6080
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 4.9481348493915816e-05,
+      "loss": 0.193,
+      "step": 6090
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 4.947137442649112e-05,
+      "loss": 0.1823,
+      "step": 6100
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 4.946140035906643e-05,
+      "loss": 0.2303,
+      "step": 6110
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.9451426291641735e-05,
+      "loss": 0.1933,
+      "step": 6120
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 4.944145222421704e-05,
+      "loss": 0.2238,
+      "step": 6130
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 4.943147815679234e-05,
+      "loss": 0.1762,
+      "step": 6140
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 4.9421504089367646e-05,
+      "loss": 0.2441,
+      "step": 6150
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.941153002194295e-05,
+      "loss": 0.1548,
+      "step": 6160
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.940155595451825e-05,
+      "loss": 0.2052,
+      "step": 6170
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 4.939158188709356e-05,
+      "loss": 0.1747,
+      "step": 6180
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 4.9381607819668864e-05,
+      "loss": 0.1918,
+      "step": 6190
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 4.937163375224417e-05,
+      "loss": 0.1776,
+      "step": 6200
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.936165968481947e-05,
+      "loss": 0.2143,
+      "step": 6210
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 4.9351685617394776e-05,
+      "loss": 0.2272,
+      "step": 6220
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 4.934171154997008e-05,
+      "loss": 0.2238,
+      "step": 6230
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 4.933173748254539e-05,
+      "loss": 0.215,
+      "step": 6240
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.932176341512069e-05,
+      "loss": 0.2505,
+      "step": 6250
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.9311789347695994e-05,
+      "loss": 0.193,
+      "step": 6260
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 4.9301815280271294e-05,
+      "loss": 0.2112,
+      "step": 6270
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 4.92918412128466e-05,
+      "loss": 0.1776,
+      "step": 6280
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 4.9281867145421906e-05,
+      "loss": 0.1934,
+      "step": 6290
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.927189307799721e-05,
+      "loss": 0.1629,
+      "step": 6300
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 4.926191901057252e-05,
+      "loss": 0.2348,
+      "step": 6310
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 4.925194494314782e-05,
+      "loss": 0.2408,
+      "step": 6320
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 4.924197087572312e-05,
+      "loss": 0.1725,
+      "step": 6330
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.923199680829842e-05,
+      "loss": 0.1752,
+      "step": 6340
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 4.922202274087373e-05,
+      "loss": 0.185,
+      "step": 6350
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.9212048673449036e-05,
+      "loss": 0.2098,
+      "step": 6360
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 4.920207460602434e-05,
+      "loss": 0.2003,
+      "step": 6370
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 4.919210053859965e-05,
+      "loss": 0.2272,
+      "step": 6380
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.918212647117495e-05,
+      "loss": 0.2392,
+      "step": 6390
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 4.917215240375025e-05,
+      "loss": 0.2252,
+      "step": 6400
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 4.916217833632555e-05,
+      "loss": 0.1832,
+      "step": 6410
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.915220426890086e-05,
+      "loss": 0.1667,
+      "step": 6420
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 4.9142230201476165e-05,
+      "loss": 0.2217,
+      "step": 6430
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 4.913225613405147e-05,
+      "loss": 0.1898,
+      "step": 6440
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 4.912228206662677e-05,
+      "loss": 0.2141,
+      "step": 6450
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.911230799920208e-05,
+      "loss": 0.1524,
+      "step": 6460
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.9102333931777383e-05,
+      "loss": 0.1946,
+      "step": 6470
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 4.909235986435268e-05,
+      "loss": 0.2343,
+      "step": 6480
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 4.908238579692799e-05,
+      "loss": 0.1926,
+      "step": 6490
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 4.9072411729503295e-05,
+      "loss": 0.1683,
+      "step": 6500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.9062437662078595e-05,
+      "loss": 0.188,
+      "step": 6510
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 4.90524635946539e-05,
+      "loss": 0.2279,
+      "step": 6520
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 4.904248952722921e-05,
+      "loss": 0.1946,
+      "step": 6530
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 4.903251545980451e-05,
+      "loss": 0.1914,
+      "step": 6540
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.902254139237982e-05,
+      "loss": 0.1413,
+      "step": 6550
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.901256732495512e-05,
+      "loss": 0.2083,
+      "step": 6560
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 4.900259325753042e-05,
+      "loss": 0.1975,
+      "step": 6570
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 4.8992619190105724e-05,
+      "loss": 0.1888,
+      "step": 6580
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 4.898264512268103e-05,
+      "loss": 0.1991,
+      "step": 6590
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.897267105525634e-05,
+      "loss": 0.1945,
+      "step": 6600
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 4.896269698783164e-05,
+      "loss": 0.1908,
+      "step": 6610
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 4.895272292040695e-05,
+      "loss": 0.1826,
+      "step": 6620
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 4.894274885298225e-05,
+      "loss": 0.1633,
+      "step": 6630
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.893277478555755e-05,
+      "loss": 0.2038,
+      "step": 6640
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.8922800718132854e-05,
+      "loss": 0.1872,
+      "step": 6650
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 4.891282665070816e-05,
+      "loss": 0.1859,
+      "step": 6660
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 4.890285258328347e-05,
+      "loss": 0.2032,
+      "step": 6670
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 4.889287851585877e-05,
+      "loss": 0.2241,
+      "step": 6680
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.9605292666517157
+      },
+      "eval_f1": {
+        "f1": 0.9584812990028835
+      },
+      "eval_loss": 0.11326823383569717,
+      "eval_precision": {
+        "precision": 0.9604748305069212
+      },
+      "eval_recall": {
+        "recall": 0.9567171466267748
+      },
+      "eval_runtime": 164.5128,
+      "eval_samples_per_second": 108.417,
+      "eval_steps_per_second": 6.778,
+      "step": 6688
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.888290444843407e-05,
+      "loss": 0.2208,
+      "step": 6690
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 4.887293038100938e-05,
+      "loss": 0.1563,
+      "step": 6700
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 4.8862956313584685e-05,
+      "loss": 0.2045,
+      "step": 6710
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 4.8852982246159984e-05,
+      "loss": 0.2119,
+      "step": 6720
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.884300817873529e-05,
+      "loss": 0.1606,
+      "step": 6730
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 4.8833034111310596e-05,
+      "loss": 0.1847,
+      "step": 6740
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.8823060043885896e-05,
+      "loss": 0.1676,
+      "step": 6750
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 4.88130859764612e-05,
+      "loss": 0.1854,
+      "step": 6760
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 4.880311190903651e-05,
+      "loss": 0.2056,
+      "step": 6770
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.8793137841611814e-05,
+      "loss": 0.1767,
+      "step": 6780
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 4.878316377418712e-05,
+      "loss": 0.1851,
+      "step": 6790
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 4.877318970676242e-05,
+      "loss": 0.183,
+      "step": 6800
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 4.876321563933772e-05,
+      "loss": 0.1944,
+      "step": 6810
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.8753241571913026e-05,
+      "loss": 0.1925,
+      "step": 6820
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 4.874326750448833e-05,
+      "loss": 0.2301,
+      "step": 6830
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.873329343706364e-05,
+      "loss": 0.1632,
+      "step": 6840
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 4.8723319369638944e-05,
+      "loss": 0.1924,
+      "step": 6850
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 4.871334530221425e-05,
+      "loss": 0.1527,
+      "step": 6860
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.870337123478955e-05,
+      "loss": 0.2015,
+      "step": 6870
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 4.869339716736485e-05,
+      "loss": 0.2006,
+      "step": 6880
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 4.8683423099940155e-05,
+      "loss": 0.1515,
+      "step": 6890
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 4.867344903251546e-05,
+      "loss": 0.1969,
+      "step": 6900
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.866347496509077e-05,
+      "loss": 0.2036,
+      "step": 6910
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 4.8653500897666074e-05,
+      "loss": 0.1788,
+      "step": 6920
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 4.8643526830241373e-05,
+      "loss": 0.1581,
+      "step": 6930
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.863355276281668e-05,
+      "loss": 0.1733,
+      "step": 6940
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 4.8623578695391986e-05,
+      "loss": 0.1951,
+      "step": 6950
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.8613604627967285e-05,
+      "loss": 0.1607,
+      "step": 6960
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 4.860363056054259e-05,
+      "loss": 0.2152,
+      "step": 6970
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 4.85936564931179e-05,
+      "loss": 0.1884,
+      "step": 6980
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.85836824256932e-05,
+      "loss": 0.196,
+      "step": 6990
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 4.85737083582685e-05,
+      "loss": 0.1845,
+      "step": 7000
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 4.856373429084381e-05,
+      "loss": 0.1394,
+      "step": 7010
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 4.8553760223419116e-05,
+      "loss": 0.2307,
+      "step": 7020
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.854378615599442e-05,
+      "loss": 0.142,
+      "step": 7030
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.853381208856972e-05,
+      "loss": 0.1859,
+      "step": 7040
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 4.852383802114502e-05,
+      "loss": 0.2221,
+      "step": 7050
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 4.851386395372033e-05,
+      "loss": 0.1906,
+      "step": 7060
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 4.850388988629563e-05,
+      "loss": 0.1679,
+      "step": 7070
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.849391581887094e-05,
+      "loss": 0.1754,
+      "step": 7080
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 4.8483941751446245e-05,
+      "loss": 0.1901,
+      "step": 7090
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 4.847396768402155e-05,
+      "loss": 0.1671,
+      "step": 7100
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 4.846399361659685e-05,
+      "loss": 0.2069,
+      "step": 7110
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.845401954917215e-05,
+      "loss": 0.1864,
+      "step": 7120
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 4.8444045481747457e-05,
+      "loss": 0.1979,
+      "step": 7130
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.843407141432276e-05,
+      "loss": 0.1953,
+      "step": 7140
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 4.842409734689807e-05,
+      "loss": 0.2274,
+      "step": 7150
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 4.8414123279473375e-05,
+      "loss": 0.1881,
+      "step": 7160
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.8404149212048675e-05,
+      "loss": 0.213,
+      "step": 7170
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 4.839417514462398e-05,
+      "loss": 0.1931,
+      "step": 7180
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 4.838420107719928e-05,
+      "loss": 0.191,
+      "step": 7190
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 4.8374227009774586e-05,
+      "loss": 0.1893,
+      "step": 7200
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.836425294234989e-05,
+      "loss": 0.1639,
+      "step": 7210
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 4.83542788749252e-05,
+      "loss": 0.1494,
+      "step": 7220
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.83443048075005e-05,
+      "loss": 0.2012,
+      "step": 7230
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 4.8334330740075804e-05,
+      "loss": 0.246,
+      "step": 7240
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 4.832435667265111e-05,
+      "loss": 0.1742,
+      "step": 7250
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.831438260522642e-05,
+      "loss": 0.1408,
+      "step": 7260
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 4.8304408537801716e-05,
+      "loss": 0.1961,
+      "step": 7270
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 4.829443447037702e-05,
+      "loss": 0.2584,
+      "step": 7280
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 4.828446040295232e-05,
+      "loss": 0.1761,
+      "step": 7290
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.827448633552763e-05,
+      "loss": 0.2035,
+      "step": 7300
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 4.8264512268102934e-05,
+      "loss": 0.2017,
+      "step": 7310
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 4.825453820067824e-05,
+      "loss": 0.1807,
+      "step": 7320
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.8244564133253547e-05,
+      "loss": 0.1653,
+      "step": 7330
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 4.823459006582885e-05,
+      "loss": 0.1892,
+      "step": 7340
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.822461599840415e-05,
+      "loss": 0.1663,
+      "step": 7350
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 4.821464193097945e-05,
+      "loss": 0.1638,
+      "step": 7360
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 4.820466786355476e-05,
+      "loss": 0.2148,
+      "step": 7370
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 4.8194693796130064e-05,
+      "loss": 0.1801,
+      "step": 7380
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.818471972870537e-05,
+      "loss": 0.2135,
+      "step": 7390
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 4.8174745661280676e-05,
+      "loss": 0.1737,
+      "step": 7400
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 4.8164771593855976e-05,
+      "loss": 0.1583,
+      "step": 7410
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 4.815479752643128e-05,
+      "loss": 0.1403,
+      "step": 7420
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.814482345900658e-05,
+      "loss": 0.223,
+      "step": 7430
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.813484939158189e-05,
+      "loss": 0.1794,
+      "step": 7440
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 4.8124875324157194e-05,
+      "loss": 0.182,
+      "step": 7450
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 4.81149012567325e-05,
+      "loss": 0.1429,
+      "step": 7460
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 4.81049271893078e-05,
+      "loss": 0.1657,
+      "step": 7470
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.8094953121883106e-05,
+      "loss": 0.1623,
+      "step": 7480
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 4.808497905445841e-05,
+      "loss": 0.2058,
+      "step": 7490
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 4.807500498703372e-05,
+      "loss": 0.1683,
+      "step": 7500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 4.806503091960902e-05,
+      "loss": 0.1531,
+      "step": 7510
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.8055056852184324e-05,
+      "loss": 0.1793,
+      "step": 7520
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.804508278475963e-05,
+      "loss": 0.1785,
+      "step": 7530
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 4.803510871733493e-05,
+      "loss": 0.2247,
+      "step": 7540
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.8025134649910235e-05,
+      "loss": 0.1673,
+      "step": 7550
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.801516058248554e-05,
+      "loss": 0.2173,
+      "step": 7560
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.800518651506085e-05,
+      "loss": 0.1675,
+      "step": 7570
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.7995212447636154e-05,
+      "loss": 0.2222,
+      "step": 7580
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.798523838021145e-05,
+      "loss": 0.2106,
+      "step": 7590
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.797526431278675e-05,
+      "loss": 0.177,
+      "step": 7600
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.796529024536206e-05,
+      "loss": 0.1841,
+      "step": 7610
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.7955316177937365e-05,
+      "loss": 0.16,
+      "step": 7620
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.794534211051267e-05,
+      "loss": 0.1899,
+      "step": 7630
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.793536804308798e-05,
+      "loss": 0.1864,
+      "step": 7640
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.792539397566328e-05,
+      "loss": 0.137,
+      "step": 7650
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.791541990823858e-05,
+      "loss": 0.1933,
+      "step": 7660
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.790544584081388e-05,
+      "loss": 0.1269,
+      "step": 7670
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.789547177338919e-05,
+      "loss": 0.2185,
+      "step": 7680
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.7885497705964495e-05,
+      "loss": 0.1833,
+      "step": 7690
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.78755236385398e-05,
+      "loss": 0.1457,
+      "step": 7700
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.78655495711151e-05,
+      "loss": 0.1443,
+      "step": 7710
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.785557550369041e-05,
+      "loss": 0.1887,
+      "step": 7720
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.784560143626571e-05,
+      "loss": 0.1302,
+      "step": 7730
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.783562736884102e-05,
+      "loss": 0.1741,
+      "step": 7740
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.782565330141632e-05,
+      "loss": 0.1816,
+      "step": 7750
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.7815679233991625e-05,
+      "loss": 0.1846,
+      "step": 7760
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.780570516656693e-05,
+      "loss": 0.1754,
+      "step": 7770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.779573109914223e-05,
+      "loss": 0.1595,
+      "step": 7780
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.7785757031717536e-05,
+      "loss": 0.185,
+      "step": 7790
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.777578296429284e-05,
+      "loss": 0.2375,
+      "step": 7800
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.9638932496075353
+      },
+      "eval_f1": {
+        "f1": 0.9615520166695122
+      },
+      "eval_loss": 0.11376180499792099,
+      "eval_precision": {
+        "precision": 0.9628392910079963
+      },
+      "eval_recall": {
+        "recall": 0.9605232709871921
+      },
+      "eval_runtime": 164.872,
+      "eval_samples_per_second": 108.181,
+      "eval_steps_per_second": 6.763,
+      "step": 7803
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.776580889686815e-05,
+      "loss": 0.1583,
+      "step": 7810
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.775583482944345e-05,
+      "loss": 0.1386,
+      "step": 7820
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.7745860762018754e-05,
+      "loss": 0.1907,
+      "step": 7830
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.7735886694594054e-05,
+      "loss": 0.1828,
+      "step": 7840
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.772591262716936e-05,
+      "loss": 0.2087,
+      "step": 7850
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.7715938559744666e-05,
+      "loss": 0.1745,
+      "step": 7860
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.770596449231997e-05,
+      "loss": 0.1646,
+      "step": 7870
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.769599042489528e-05,
+      "loss": 0.1477,
+      "step": 7880
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.768601635747058e-05,
+      "loss": 0.1652,
+      "step": 7890
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.767604229004588e-05,
+      "loss": 0.1279,
+      "step": 7900
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.7666068222621184e-05,
+      "loss": 0.1598,
+      "step": 7910
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.765609415519649e-05,
+      "loss": 0.1724,
+      "step": 7920
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.7646120087771796e-05,
+      "loss": 0.1433,
+      "step": 7930
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.76361460203471e-05,
+      "loss": 0.1708,
+      "step": 7940
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.76261719529224e-05,
+      "loss": 0.2265,
+      "step": 7950
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.761619788549771e-05,
+      "loss": 0.1611,
+      "step": 7960
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 4.7606223818073014e-05,
+      "loss": 0.1753,
+      "step": 7970
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 4.7596249750648313e-05,
+      "loss": 0.1625,
+      "step": 7980
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 4.758627568322362e-05,
+      "loss": 0.1698,
+      "step": 7990
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.7576301615798926e-05,
+      "loss": 0.178,
+      "step": 8000
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.756632754837423e-05,
+      "loss": 0.1586,
+      "step": 8010
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 4.755635348094953e-05,
+      "loss": 0.1753,
+      "step": 8020
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 4.754637941352484e-05,
+      "loss": 0.1779,
+      "step": 8030
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 4.7536405346100144e-05,
+      "loss": 0.1448,
+      "step": 8040
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.752643127867545e-05,
+      "loss": 0.1181,
+      "step": 8050
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 4.751645721125075e-05,
+      "loss": 0.1578,
+      "step": 8060
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 4.7506483143826056e-05,
+      "loss": 0.1995,
+      "step": 8070
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.7496509076401355e-05,
+      "loss": 0.1534,
+      "step": 8080
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 4.748653500897666e-05,
+      "loss": 0.2371,
+      "step": 8090
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 4.747656094155197e-05,
+      "loss": 0.1536,
+      "step": 8100
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.7466586874127274e-05,
+      "loss": 0.1394,
+      "step": 8110
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 4.745661280670258e-05,
+      "loss": 0.1179,
+      "step": 8120
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.744663873927788e-05,
+      "loss": 0.1618,
+      "step": 8130
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 4.743666467185318e-05,
+      "loss": 0.2594,
+      "step": 8140
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 4.7426690604428485e-05,
+      "loss": 0.1362,
+      "step": 8150
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 4.741671653700379e-05,
+      "loss": 0.1336,
+      "step": 8160
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.74067424695791e-05,
+      "loss": 0.1538,
+      "step": 8170
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 4.7396768402154403e-05,
+      "loss": 0.1864,
+      "step": 8180
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 4.73867943347297e-05,
+      "loss": 0.159,
+      "step": 8190
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.737682026730501e-05,
+      "loss": 0.1712,
+      "step": 8200
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 4.7366846199880315e-05,
+      "loss": 0.1728,
+      "step": 8210
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.7356872132455615e-05,
+      "loss": 0.2012,
+      "step": 8220
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 4.734689806503092e-05,
+      "loss": 0.1781,
+      "step": 8230
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 4.733692399760623e-05,
+      "loss": 0.1474,
+      "step": 8240
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 4.732694993018153e-05,
+      "loss": 0.1959,
+      "step": 8250
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.731697586275683e-05,
+      "loss": 0.1493,
+      "step": 8260
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 4.730700179533214e-05,
+      "loss": 0.1826,
+      "step": 8270
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 4.7297027727907445e-05,
+      "loss": 0.187,
+      "step": 8280
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 4.728705366048275e-05,
+      "loss": 0.1651,
+      "step": 8290
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.727707959305805e-05,
+      "loss": 0.1475,
+      "step": 8300
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.726710552563336e-05,
+      "loss": 0.153,
+      "step": 8310
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 4.7257131458208656e-05,
+      "loss": 0.1487,
+      "step": 8320
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 4.724715739078396e-05,
+      "loss": 0.1971,
+      "step": 8330
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 4.723718332335927e-05,
+      "loss": 0.173,
+      "step": 8340
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.7227209255934575e-05,
+      "loss": 0.1757,
+      "step": 8350
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 4.721723518850988e-05,
+      "loss": 0.1457,
+      "step": 8360
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 4.720726112108518e-05,
+      "loss": 0.1425,
+      "step": 8370
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 4.719728705366048e-05,
+      "loss": 0.1448,
+      "step": 8380
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.7187312986235786e-05,
+      "loss": 0.1428,
+      "step": 8390
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.717733891881109e-05,
+      "loss": 0.2053,
+      "step": 8400
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 4.71673648513864e-05,
+      "loss": 0.1575,
+      "step": 8410
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 4.7157390783961705e-05,
+      "loss": 0.1386,
+      "step": 8420
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 4.7147416716537004e-05,
+      "loss": 0.1825,
+      "step": 8430
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.713744264911231e-05,
+      "loss": 0.1415,
+      "step": 8440
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 4.7127468581687616e-05,
+      "loss": 0.1635,
+      "step": 8450
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 4.7117494514262916e-05,
+      "loss": 0.1666,
+      "step": 8460
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 4.710752044683822e-05,
+      "loss": 0.1335,
+      "step": 8470
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.709754637941353e-05,
+      "loss": 0.1482,
+      "step": 8480
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 4.7087572311988834e-05,
+      "loss": 0.1565,
+      "step": 8490
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.7077598244564134e-05,
+      "loss": 0.1826,
+      "step": 8500
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 4.706762417713944e-05,
+      "loss": 0.1564,
+      "step": 8510
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 4.7057650109714746e-05,
+      "loss": 0.1847,
+      "step": 8520
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.704767604229005e-05,
+      "loss": 0.126,
+      "step": 8530
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 4.703770197486535e-05,
+      "loss": 0.1683,
+      "step": 8540
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 4.702772790744066e-05,
+      "loss": 0.1469,
+      "step": 8550
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 4.701775384001596e-05,
+      "loss": 0.2163,
+      "step": 8560
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.7007779772591264e-05,
+      "loss": 0.1879,
+      "step": 8570
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 4.699780570516657e-05,
+      "loss": 0.1194,
+      "step": 8580
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.6987831637741876e-05,
+      "loss": 0.1569,
+      "step": 8590
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 4.697785757031718e-05,
+      "loss": 0.1821,
+      "step": 8600
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 4.696788350289248e-05,
+      "loss": 0.1091,
+      "step": 8610
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.695790943546778e-05,
+      "loss": 0.1731,
+      "step": 8620
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 4.694793536804309e-05,
+      "loss": 0.1228,
+      "step": 8630
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 4.693796130061839e-05,
+      "loss": 0.1568,
+      "step": 8640
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.69279872331937e-05,
+      "loss": 0.1689,
+      "step": 8650
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 4.6918013165769006e-05,
+      "loss": 0.1513,
+      "step": 8660
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 4.690803909834431e-05,
+      "loss": 0.1785,
+      "step": 8670
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 4.689806503091961e-05,
+      "loss": 0.1706,
+      "step": 8680
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.688809096349491e-05,
+      "loss": 0.1457,
+      "step": 8690
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.687811689607022e-05,
+      "loss": 0.1866,
+      "step": 8700
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 4.686814282864552e-05,
+      "loss": 0.2145,
+      "step": 8710
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 4.685816876122083e-05,
+      "loss": 0.1558,
+      "step": 8720
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 4.6848194693796136e-05,
+      "loss": 0.178,
+      "step": 8730
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.6838220626371435e-05,
+      "loss": 0.1722,
+      "step": 8740
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 4.682824655894674e-05,
+      "loss": 0.1616,
+      "step": 8750
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 4.681827249152205e-05,
+      "loss": 0.1963,
+      "step": 8760
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 4.680829842409735e-05,
+      "loss": 0.1461,
+      "step": 8770
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.679832435667265e-05,
+      "loss": 0.1803,
+      "step": 8780
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.678835028924796e-05,
+      "loss": 0.1707,
+      "step": 8790
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 4.677837622182326e-05,
+      "loss": 0.1378,
+      "step": 8800
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 4.6768402154398565e-05,
+      "loss": 0.1328,
+      "step": 8810
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 4.675842808697387e-05,
+      "loss": 0.1801,
+      "step": 8820
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.674845401954918e-05,
+      "loss": 0.1892,
+      "step": 8830
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 4.673847995212448e-05,
+      "loss": 0.1681,
+      "step": 8840
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 4.672850588469978e-05,
+      "loss": 0.1814,
+      "step": 8850
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 4.671853181727508e-05,
+      "loss": 0.1779,
+      "step": 8860
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.670855774985039e-05,
+      "loss": 0.1453,
+      "step": 8870
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.6698583682425695e-05,
+      "loss": 0.1376,
+      "step": 8880
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 4.6688609615001e-05,
+      "loss": 0.1833,
+      "step": 8890
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 4.667863554757631e-05,
+      "loss": 0.157,
+      "step": 8900
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 4.666866148015161e-05,
+      "loss": 0.1946,
+      "step": 8910
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.9682103610675039
+      },
+      "eval_f1": {
+        "f1": 0.9665726403969513
+      },
+      "eval_loss": 0.09346462041139603,
+      "eval_precision": {
+        "precision": 0.9677788948564787
+      },
+      "eval_recall": {
+        "recall": 0.9654971437363139
+      },
+      "eval_runtime": 230.3224,
+      "eval_samples_per_second": 77.439,
+      "eval_steps_per_second": 4.841,
+      "step": 8918
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.665868741272691e-05,
+      "loss": 0.13,
+      "step": 8920
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 4.664871334530221e-05,
+      "loss": 0.1427,
+      "step": 8930
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 4.663873927787752e-05,
+      "loss": 0.134,
+      "step": 8940
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 4.6628765210452824e-05,
+      "loss": 0.1438,
+      "step": 8950
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.661879114302813e-05,
+      "loss": 0.1588,
+      "step": 8960
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 4.660881707560344e-05,
+      "loss": 0.1759,
+      "step": 8970
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.6598843008178736e-05,
+      "loss": 0.1587,
+      "step": 8980
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 4.658886894075404e-05,
+      "loss": 0.1554,
+      "step": 8990
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 4.657889487332935e-05,
+      "loss": 0.1699,
+      "step": 9000
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.656892080590465e-05,
+      "loss": 0.1696,
+      "step": 9010
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 4.6558946738479954e-05,
+      "loss": 0.1248,
+      "step": 9020
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 4.654897267105526e-05,
+      "loss": 0.1347,
+      "step": 9030
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 4.653899860363056e-05,
+      "loss": 0.1782,
+      "step": 9040
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.6529024536205866e-05,
+      "loss": 0.1303,
+      "step": 9050
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 4.651905046878117e-05,
+      "loss": 0.1749,
+      "step": 9060
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 4.650907640135648e-05,
+      "loss": 0.1453,
+      "step": 9070
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.6499102333931784e-05,
+      "loss": 0.1328,
+      "step": 9080
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 4.6489128266507084e-05,
+      "loss": 0.1688,
+      "step": 9090
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.647915419908238e-05,
+      "loss": 0.1408,
+      "step": 9100
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 4.646918013165769e-05,
+      "loss": 0.1646,
+      "step": 9110
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 4.6459206064232996e-05,
+      "loss": 0.1303,
+      "step": 9120
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 4.64492319968083e-05,
+      "loss": 0.1341,
+      "step": 9130
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.643925792938361e-05,
+      "loss": 0.1239,
+      "step": 9140
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 4.6429283861958914e-05,
+      "loss": 0.2092,
+      "step": 9150
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 4.6419309794534214e-05,
+      "loss": 0.2008,
+      "step": 9160
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 4.640933572710951e-05,
+      "loss": 0.1682,
+      "step": 9170
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.639936165968482e-05,
+      "loss": 0.1514,
+      "step": 9180
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.6389387592260125e-05,
+      "loss": 0.1572,
+      "step": 9190
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 4.637941352483543e-05,
+      "loss": 0.1411,
+      "step": 9200
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 4.636943945741074e-05,
+      "loss": 0.1377,
+      "step": 9210
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.635946538998604e-05,
+      "loss": 0.1551,
+      "step": 9220
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 4.6349491322561343e-05,
+      "loss": 0.1678,
+      "step": 9230
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 4.633951725513665e-05,
+      "loss": 0.1507,
+      "step": 9240
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 4.632954318771195e-05,
+      "loss": 0.1749,
+      "step": 9250
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.6319569120287255e-05,
+      "loss": 0.1652,
+      "step": 9260
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.630959505286256e-05,
+      "loss": 0.1204,
+      "step": 9270
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 4.629962098543786e-05,
+      "loss": 0.1848,
+      "step": 9280
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 4.628964691801317e-05,
+      "loss": 0.1622,
+      "step": 9290
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 4.627967285058847e-05,
+      "loss": 0.1577,
+      "step": 9300
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.626969878316378e-05,
+      "loss": 0.1523,
+      "step": 9310
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 4.625972471573908e-05,
+      "loss": 0.1554,
+      "step": 9320
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 4.6249750648314385e-05,
+      "loss": 0.1719,
+      "step": 9330
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 4.6239776580889684e-05,
+      "loss": 0.1434,
+      "step": 9340
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.622980251346499e-05,
+      "loss": 0.1778,
+      "step": 9350
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 4.62198284460403e-05,
+      "loss": 0.1408,
+      "step": 9360
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.62098543786156e-05,
+      "loss": 0.1348,
+      "step": 9370
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 4.619988031119091e-05,
+      "loss": 0.1565,
+      "step": 9380
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 4.6189906243766215e-05,
+      "loss": 0.1757,
+      "step": 9390
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 4.6179932176341515e-05,
+      "loss": 0.1481,
+      "step": 9400
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 4.6169958108916814e-05,
+      "loss": 0.1581,
+      "step": 9410
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 4.615998404149212e-05,
+      "loss": 0.1428,
+      "step": 9420
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 4.615000997406743e-05,
+      "loss": 0.1975,
+      "step": 9430
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 4.614003590664273e-05,
+      "loss": 0.1243,
+      "step": 9440
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 4.613006183921804e-05,
+      "loss": 0.1247,
+      "step": 9450
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 4.612008777179334e-05,
+      "loss": 0.1407,
+      "step": 9460
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.6110113704368645e-05,
+      "loss": 0.2017,
+      "step": 9470
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 4.6100139636943944e-05,
+      "loss": 0.0993,
+      "step": 9480
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.609016556951925e-05,
+      "loss": 0.1301,
+      "step": 9490
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.6080191502094556e-05,
+      "loss": 0.1576,
+      "step": 9500
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.607021743466986e-05,
+      "loss": 0.1534,
+      "step": 9510
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 4.606024336724516e-05,
+      "loss": 0.1482,
+      "step": 9520
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.605026929982047e-05,
+      "loss": 0.1517,
+      "step": 9530
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.6040295232395774e-05,
+      "loss": 0.1516,
+      "step": 9540
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.603032116497108e-05,
+      "loss": 0.1502,
+      "step": 9550
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.602034709754638e-05,
+      "loss": 0.1608,
+      "step": 9560
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.6010373030121686e-05,
+      "loss": 0.1548,
+      "step": 9570
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 4.6000398962696986e-05,
+      "loss": 0.1462,
+      "step": 9580
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.599042489527229e-05,
+      "loss": 0.1413,
+      "step": 9590
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.59804508278476e-05,
+      "loss": 0.169,
+      "step": 9600
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.5970476760422904e-05,
+      "loss": 0.1272,
+      "step": 9610
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 4.596050269299821e-05,
+      "loss": 0.182,
+      "step": 9620
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.5950528625573517e-05,
+      "loss": 0.1559,
+      "step": 9630
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.5940554558148816e-05,
+      "loss": 0.1296,
+      "step": 9640
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.5930580490724115e-05,
+      "loss": 0.1308,
+      "step": 9650
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.592060642329942e-05,
+      "loss": 0.1376,
+      "step": 9660
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.591063235587473e-05,
+      "loss": 0.1521,
+      "step": 9670
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 4.5900658288450034e-05,
+      "loss": 0.1433,
+      "step": 9680
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.589068422102534e-05,
+      "loss": 0.1587,
+      "step": 9690
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.588071015360064e-05,
+      "loss": 0.1157,
+      "step": 9700
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.5870736086175946e-05,
+      "loss": 0.1273,
+      "step": 9710
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.5860762018751245e-05,
+      "loss": 0.1353,
+      "step": 9720
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 4.585078795132655e-05,
+      "loss": 0.1353,
+      "step": 9730
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.584081388390186e-05,
+      "loss": 0.1622,
+      "step": 9740
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.5830839816477164e-05,
+      "loss": 0.1571,
+      "step": 9750
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.582086574905246e-05,
+      "loss": 0.113,
+      "step": 9760
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.581089168162777e-05,
+      "loss": 0.1695,
+      "step": 9770
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.5800917614203076e-05,
+      "loss": 0.177,
+      "step": 9780
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 4.579094354677838e-05,
+      "loss": 0.1434,
+      "step": 9790
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.578096947935368e-05,
+      "loss": 0.1725,
+      "step": 9800
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 4.577099541192899e-05,
+      "loss": 0.1556,
+      "step": 9810
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 4.576102134450429e-05,
+      "loss": 0.1525,
+      "step": 9820
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 4.575104727707959e-05,
+      "loss": 0.1449,
+      "step": 9830
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 4.57410732096549e-05,
+      "loss": 0.1502,
+      "step": 9840
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 4.5731099142230205e-05,
+      "loss": 0.1659,
+      "step": 9850
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.572112507480551e-05,
+      "loss": 0.1629,
+      "step": 9860
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 4.571115100738082e-05,
+      "loss": 0.1381,
+      "step": 9870
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 4.570117693995612e-05,
+      "loss": 0.1375,
+      "step": 9880
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 4.5691202872531417e-05,
+      "loss": 0.1237,
+      "step": 9890
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 4.568122880510672e-05,
+      "loss": 0.1987,
+      "step": 9900
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 4.567125473768203e-05,
+      "loss": 0.1051,
+      "step": 9910
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 4.5661280670257335e-05,
+      "loss": 0.1314,
+      "step": 9920
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 4.565130660283264e-05,
+      "loss": 0.1568,
+      "step": 9930
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 4.564133253540794e-05,
+      "loss": 0.1273,
+      "step": 9940
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.563135846798325e-05,
+      "loss": 0.1329,
+      "step": 9950
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 4.5621384400558546e-05,
+      "loss": 0.1381,
+      "step": 9960
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 4.561141033313385e-05,
+      "loss": 0.1637,
+      "step": 9970
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 4.560143626570916e-05,
+      "loss": 0.137,
+      "step": 9980
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 4.5591462198284465e-05,
+      "loss": 0.137,
+      "step": 9990
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 4.5581488130859764e-05,
+      "loss": 0.1327,
+      "step": 10000
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 4.557151406343507e-05,
+      "loss": 0.1607,
+      "step": 10010
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 4.556153999601038e-05,
+      "loss": 0.1448,
+      "step": 10020
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.555156592858568e-05,
+      "loss": 0.1513,
+      "step": 10030
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.967593630858937
+      },
+      "eval_f1": {
+        "f1": 0.9657913168767667
+      },
+      "eval_loss": 0.09078183770179749,
+      "eval_precision": {
+        "precision": 0.9657840712769861
+      },
+      "eval_recall": {
+        "recall": 0.9659100270116727
+      },
+      "eval_runtime": 161.7027,
+      "eval_samples_per_second": 110.301,
+      "eval_steps_per_second": 6.895,
+      "step": 10032
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.554159186116098e-05,
+      "loss": 0.1404,
+      "step": 10040
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.553161779373629e-05,
+      "loss": 0.127,
+      "step": 10050
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.5521643726311595e-05,
+      "loss": 0.1565,
+      "step": 10060
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.5511669658886894e-05,
+      "loss": 0.1519,
+      "step": 10070
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.55016955914622e-05,
+      "loss": 0.1463,
+      "step": 10080
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.5491721524037507e-05,
+      "loss": 0.1424,
+      "step": 10090
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.548174745661281e-05,
+      "loss": 0.1455,
+      "step": 10100
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.547177338918811e-05,
+      "loss": 0.1326,
+      "step": 10110
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.546179932176342e-05,
+      "loss": 0.1404,
+      "step": 10120
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.545182525433872e-05,
+      "loss": 0.1314,
+      "step": 10130
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.5441851186914024e-05,
+      "loss": 0.138,
+      "step": 10140
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.543187711948933e-05,
+      "loss": 0.1371,
+      "step": 10150
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.5421903052064636e-05,
+      "loss": 0.1742,
+      "step": 10160
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.541192898463994e-05,
+      "loss": 0.1123,
+      "step": 10170
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.540195491721524e-05,
+      "loss": 0.11,
+      "step": 10180
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.539198084979054e-05,
+      "loss": 0.1648,
+      "step": 10190
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.538200678236585e-05,
+      "loss": 0.1932,
+      "step": 10200
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.5372032714941154e-05,
+      "loss": 0.1307,
+      "step": 10210
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.536205864751646e-05,
+      "loss": 0.1472,
+      "step": 10220
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.5352084580091766e-05,
+      "loss": 0.1158,
+      "step": 10230
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.5342110512667066e-05,
+      "loss": 0.127,
+      "step": 10240
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.533213644524237e-05,
+      "loss": 0.1478,
+      "step": 10250
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.532216237781768e-05,
+      "loss": 0.1328,
+      "step": 10260
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.531218831039298e-05,
+      "loss": 0.1285,
+      "step": 10270
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.5302214242968284e-05,
+      "loss": 0.1407,
+      "step": 10280
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.529224017554359e-05,
+      "loss": 0.1482,
+      "step": 10290
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.5282266108118896e-05,
+      "loss": 0.1315,
+      "step": 10300
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.5272292040694195e-05,
+      "loss": 0.1114,
+      "step": 10310
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.52623179732695e-05,
+      "loss": 0.1132,
+      "step": 10320
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.525234390584481e-05,
+      "loss": 0.1614,
+      "step": 10330
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.5242369838420114e-05,
+      "loss": 0.0982,
+      "step": 10340
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 4.523239577099541e-05,
+      "loss": 0.116,
+      "step": 10350
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 4.522242170357072e-05,
+      "loss": 0.1536,
+      "step": 10360
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 4.521244763614602e-05,
+      "loss": 0.1613,
+      "step": 10370
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 4.5202473568721325e-05,
+      "loss": 0.1352,
+      "step": 10380
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 4.519249950129663e-05,
+      "loss": 0.1123,
+      "step": 10390
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 4.518252543387194e-05,
+      "loss": 0.1556,
+      "step": 10400
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 4.5172551366447244e-05,
+      "loss": 0.1508,
+      "step": 10410
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 4.516257729902254e-05,
+      "loss": 0.1321,
+      "step": 10420
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 4.515260323159784e-05,
+      "loss": 0.1153,
+      "step": 10430
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.514262916417315e-05,
+      "loss": 0.1233,
+      "step": 10440
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 4.5132655096748455e-05,
+      "loss": 0.149,
+      "step": 10450
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.512268102932376e-05,
+      "loss": 0.1475,
+      "step": 10460
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 4.511270696189907e-05,
+      "loss": 0.1516,
+      "step": 10470
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 4.510273289447437e-05,
+      "loss": 0.1501,
+      "step": 10480
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 4.509275882704967e-05,
+      "loss": 0.1466,
+      "step": 10490
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 4.508278475962498e-05,
+      "loss": 0.1111,
+      "step": 10500
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 4.507281069220028e-05,
+      "loss": 0.1095,
+      "step": 10510
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 4.5062836624775585e-05,
+      "loss": 0.1676,
+      "step": 10520
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 4.505286255735089e-05,
+      "loss": 0.1508,
+      "step": 10530
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.50428884899262e-05,
+      "loss": 0.1371,
+      "step": 10540
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 4.5032914422501496e-05,
+      "loss": 0.1374,
+      "step": 10550
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 4.50229403550768e-05,
+      "loss": 0.1284,
+      "step": 10560
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 4.501296628765211e-05,
+      "loss": 0.1624,
+      "step": 10570
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 4.5002992220227415e-05,
+      "loss": 0.1197,
+      "step": 10580
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 4.4993018152802714e-05,
+      "loss": 0.1362,
+      "step": 10590
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 4.498304408537802e-05,
+      "loss": 0.1399,
+      "step": 10600
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 4.497307001795332e-05,
+      "loss": 0.1787,
+      "step": 10610
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 4.4963095950528626e-05,
+      "loss": 0.1253,
+      "step": 10620
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.495312188310393e-05,
+      "loss": 0.1425,
+      "step": 10630
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 4.494314781567924e-05,
+      "loss": 0.1147,
+      "step": 10640
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 4.4933173748254545e-05,
+      "loss": 0.116,
+      "step": 10650
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 4.4923199680829844e-05,
+      "loss": 0.1059,
+      "step": 10660
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 4.4913225613405144e-05,
+      "loss": 0.1423,
+      "step": 10670
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 4.490325154598045e-05,
+      "loss": 0.1705,
+      "step": 10680
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 4.4893277478555756e-05,
+      "loss": 0.1296,
+      "step": 10690
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 4.488330341113106e-05,
+      "loss": 0.1463,
+      "step": 10700
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 4.487332934370637e-05,
+      "loss": 0.1167,
+      "step": 10710
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 4.486335527628167e-05,
+      "loss": 0.1496,
+      "step": 10720
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.4853381208856974e-05,
+      "loss": 0.1525,
+      "step": 10730
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 4.484340714143228e-05,
+      "loss": 0.1082,
+      "step": 10740
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 4.483343307400758e-05,
+      "loss": 0.1599,
+      "step": 10750
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 4.4823459006582886e-05,
+      "loss": 0.1531,
+      "step": 10760
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 4.481348493915819e-05,
+      "loss": 0.1394,
+      "step": 10770
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 4.48035108717335e-05,
+      "loss": 0.188,
+      "step": 10780
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 4.47935368043088e-05,
+      "loss": 0.119,
+      "step": 10790
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.4783562736884104e-05,
+      "loss": 0.0981,
+      "step": 10800
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.477358866945941e-05,
+      "loss": 0.139,
+      "step": 10810
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.476361460203471e-05,
+      "loss": 0.165,
+      "step": 10820
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.4753640534610016e-05,
+      "loss": 0.1459,
+      "step": 10830
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.474366646718532e-05,
+      "loss": 0.1108,
+      "step": 10840
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 4.473369239976062e-05,
+      "loss": 0.1419,
+      "step": 10850
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 4.472371833233593e-05,
+      "loss": 0.1569,
+      "step": 10860
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 4.4713744264911234e-05,
+      "loss": 0.1169,
+      "step": 10870
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 4.470377019748654e-05,
+      "loss": 0.1318,
+      "step": 10880
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 4.4693796130061846e-05,
+      "loss": 0.1554,
+      "step": 10890
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 4.4683822062637145e-05,
+      "loss": 0.1428,
+      "step": 10900
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 4.4673847995212445e-05,
+      "loss": 0.1412,
+      "step": 10910
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.466387392778775e-05,
+      "loss": 0.1275,
+      "step": 10920
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 4.465389986036306e-05,
+      "loss": 0.1336,
+      "step": 10930
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 4.4643925792938363e-05,
+      "loss": 0.1567,
+      "step": 10940
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 4.463395172551367e-05,
+      "loss": 0.1462,
+      "step": 10950
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 4.462397765808897e-05,
+      "loss": 0.1273,
+      "step": 10960
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 4.4614003590664275e-05,
+      "loss": 0.1633,
+      "step": 10970
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 4.4604029523239575e-05,
+      "loss": 0.1069,
+      "step": 10980
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.459405545581488e-05,
+      "loss": 0.1124,
+      "step": 10990
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.458408138839019e-05,
+      "loss": 0.1453,
+      "step": 11000
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 4.457410732096549e-05,
+      "loss": 0.1291,
+      "step": 11010
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.45641332535408e-05,
+      "loss": 0.1484,
+      "step": 11020
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 4.45541591861161e-05,
+      "loss": 0.1405,
+      "step": 11030
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.4544185118691405e-05,
+      "loss": 0.1353,
+      "step": 11040
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.453421105126671e-05,
+      "loss": 0.1434,
+      "step": 11050
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.452423698384201e-05,
+      "loss": 0.1495,
+      "step": 11060
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.451426291641732e-05,
+      "loss": 0.1401,
+      "step": 11070
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 4.450428884899262e-05,
+      "loss": 0.1092,
+      "step": 11080
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 4.449431478156792e-05,
+      "loss": 0.0985,
+      "step": 11090
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 4.448434071414323e-05,
+      "loss": 0.1301,
+      "step": 11100
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.4474366646718535e-05,
+      "loss": 0.1423,
+      "step": 11110
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.446439257929384e-05,
+      "loss": 0.1632,
+      "step": 11120
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.445441851186915e-05,
+      "loss": 0.1175,
+      "step": 11130
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.174,
+      "step": 11140
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.968490692980489
+      },
+      "eval_f1": {
+        "f1": 0.9669266120574112
+      },
+      "eval_loss": 0.08508118242025375,
+      "eval_precision": {
+        "precision": 0.9663934578407292
+      },
+      "eval_recall": {
+        "recall": 0.967857117571674
+      },
+      "eval_runtime": 161.4589,
+      "eval_samples_per_second": 110.468,
+      "eval_steps_per_second": 6.906,
+      "step": 11147
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 4.4434470377019746e-05,
+      "loss": 0.1638,
+      "step": 11150
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 4.442449630959505e-05,
+      "loss": 0.0967,
+      "step": 11160
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 4.441452224217036e-05,
+      "loss": 0.1045,
+      "step": 11170
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 4.4404548174745665e-05,
+      "loss": 0.1126,
+      "step": 11180
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 4.439457410732097e-05,
+      "loss": 0.134,
+      "step": 11190
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 4.438460003989628e-05,
+      "loss": 0.1202,
+      "step": 11200
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 4.4374625972471576e-05,
+      "loss": 0.1346,
+      "step": 11210
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 4.4364651905046876e-05,
+      "loss": 0.1126,
+      "step": 11220
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 4.435467783762218e-05,
+      "loss": 0.1039,
+      "step": 11230
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 4.434470377019749e-05,
+      "loss": 0.1232,
+      "step": 11240
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 4.4334729702772794e-05,
+      "loss": 0.1202,
+      "step": 11250
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 4.43247556353481e-05,
+      "loss": 0.132,
+      "step": 11260
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 4.43147815679234e-05,
+      "loss": 0.1259,
+      "step": 11270
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 4.4304807500498706e-05,
+      "loss": 0.1085,
+      "step": 11280
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.429483343307401e-05,
+      "loss": 0.174,
+      "step": 11290
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 4.428485936564931e-05,
+      "loss": 0.139,
+      "step": 11300
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 4.427488529822462e-05,
+      "loss": 0.1238,
+      "step": 11310
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 4.4264911230799924e-05,
+      "loss": 0.1045,
+      "step": 11320
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 4.4254937163375224e-05,
+      "loss": 0.1274,
+      "step": 11330
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 4.424496309595053e-05,
+      "loss": 0.1467,
+      "step": 11340
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 4.4234989028525836e-05,
+      "loss": 0.1263,
+      "step": 11350
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 4.422501496110114e-05,
+      "loss": 0.1147,
+      "step": 11360
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 4.421504089367645e-05,
+      "loss": 0.2058,
+      "step": 11370
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 4.420506682625175e-05,
+      "loss": 0.1258,
+      "step": 11380
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 4.419509275882705e-05,
+      "loss": 0.1085,
+      "step": 11390
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 4.418511869140235e-05,
+      "loss": 0.1639,
+      "step": 11400
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 4.417514462397766e-05,
+      "loss": 0.1116,
+      "step": 11410
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 4.4165170556552966e-05,
+      "loss": 0.1069,
+      "step": 11420
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 4.415519648912827e-05,
+      "loss": 0.119,
+      "step": 11430
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 4.414522242170358e-05,
+      "loss": 0.1327,
+      "step": 11440
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 4.413524835427888e-05,
+      "loss": 0.1529,
+      "step": 11450
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 4.412527428685418e-05,
+      "loss": 0.0969,
+      "step": 11460
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 4.411530021942948e-05,
+      "loss": 0.1318,
+      "step": 11470
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 4.410532615200479e-05,
+      "loss": 0.1449,
+      "step": 11480
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 4.4095352084580096e-05,
+      "loss": 0.1439,
+      "step": 11490
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 4.40853780171554e-05,
+      "loss": 0.1229,
+      "step": 11500
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 4.40754039497307e-05,
+      "loss": 0.1416,
+      "step": 11510
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 4.406542988230601e-05,
+      "loss": 0.1332,
+      "step": 11520
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 4.4055455814881314e-05,
+      "loss": 0.149,
+      "step": 11530
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 4.404548174745661e-05,
+      "loss": 0.1409,
+      "step": 11540
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 4.403550768003192e-05,
+      "loss": 0.0979,
+      "step": 11550
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 4.4025533612607225e-05,
+      "loss": 0.1358,
+      "step": 11560
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 4.4015559545182525e-05,
+      "loss": 0.149,
+      "step": 11570
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 4.400558547775783e-05,
+      "loss": 0.0962,
+      "step": 11580
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 4.399561141033314e-05,
+      "loss": 0.1386,
+      "step": 11590
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 4.398563734290844e-05,
+      "loss": 0.1107,
+      "step": 11600
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 4.397566327548374e-05,
+      "loss": 0.1564,
+      "step": 11610
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 4.396568920805905e-05,
+      "loss": 0.1251,
+      "step": 11620
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 4.395571514063435e-05,
+      "loss": 0.1858,
+      "step": 11630
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 4.3945741073209655e-05,
+      "loss": 0.147,
+      "step": 11640
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 4.393576700578496e-05,
+      "loss": 0.1441,
+      "step": 11650
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 4.392579293836027e-05,
+      "loss": 0.1232,
+      "step": 11660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 4.391581887093557e-05,
+      "loss": 0.1487,
+      "step": 11670
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 4.390584480351088e-05,
+      "loss": 0.1471,
+      "step": 11680
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 4.389587073608618e-05,
+      "loss": 0.1589,
+      "step": 11690
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 4.388589666866148e-05,
+      "loss": 0.1671,
+      "step": 11700
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 4.3875922601236784e-05,
+      "loss": 0.1178,
+      "step": 11710
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 4.386594853381209e-05,
+      "loss": 0.1029,
+      "step": 11720
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 4.38559744663874e-05,
+      "loss": 0.1494,
+      "step": 11730
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 4.38460003989627e-05,
+      "loss": 0.1306,
+      "step": 11740
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 4.3836026331538e-05,
+      "loss": 0.093,
+      "step": 11750
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 4.382605226411331e-05,
+      "loss": 0.1743,
+      "step": 11760
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 4.381607819668861e-05,
+      "loss": 0.1607,
+      "step": 11770
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 4.3806104129263914e-05,
+      "loss": 0.1262,
+      "step": 11780
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 4.379613006183922e-05,
+      "loss": 0.1613,
+      "step": 11790
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 4.3786155994414526e-05,
+      "loss": 0.1496,
+      "step": 11800
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 4.3776181926989826e-05,
+      "loss": 0.1422,
+      "step": 11810
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 4.376620785956513e-05,
+      "loss": 0.1362,
+      "step": 11820
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 4.375623379214044e-05,
+      "loss": 0.1805,
+      "step": 11830
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 4.3746259724715744e-05,
+      "loss": 0.1085,
+      "step": 11840
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 4.3736285657291044e-05,
+      "loss": 0.1533,
+      "step": 11850
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 4.372631158986635e-05,
+      "loss": 0.1261,
+      "step": 11860
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 4.371633752244165e-05,
+      "loss": 0.1178,
+      "step": 11870
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 4.3706363455016956e-05,
+      "loss": 0.1107,
+      "step": 11880
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 4.369638938759226e-05,
+      "loss": 0.1072,
+      "step": 11890
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 4.368641532016757e-05,
+      "loss": 0.1127,
+      "step": 11900
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 4.3676441252742874e-05,
+      "loss": 0.119,
+      "step": 11910
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 4.366646718531818e-05,
+      "loss": 0.1223,
+      "step": 11920
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 4.365649311789348e-05,
+      "loss": 0.1377,
+      "step": 11930
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 4.364651905046878e-05,
+      "loss": 0.0979,
+      "step": 11940
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 4.3636544983044085e-05,
+      "loss": 0.1596,
+      "step": 11950
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.362657091561939e-05,
+      "loss": 0.1282,
+      "step": 11960
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 4.36165968481947e-05,
+      "loss": 0.1166,
+      "step": 11970
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 4.3606622780770004e-05,
+      "loss": 0.1032,
+      "step": 11980
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 4.3596648713345303e-05,
+      "loss": 0.1237,
+      "step": 11990
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 4.358667464592061e-05,
+      "loss": 0.1814,
+      "step": 12000
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 4.357670057849591e-05,
+      "loss": 0.1246,
+      "step": 12010
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 4.3566726511071215e-05,
+      "loss": 0.109,
+      "step": 12020
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 4.355675244364652e-05,
+      "loss": 0.1197,
+      "step": 12030
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 4.354677837622183e-05,
+      "loss": 0.1149,
+      "step": 12040
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 4.353680430879713e-05,
+      "loss": 0.1765,
+      "step": 12050
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 4.352683024137243e-05,
+      "loss": 0.1267,
+      "step": 12060
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.351685617394774e-05,
+      "loss": 0.1052,
+      "step": 12070
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 4.3506882106523046e-05,
+      "loss": 0.1577,
+      "step": 12080
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.3496908039098345e-05,
+      "loss": 0.159,
+      "step": 12090
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.348693397167365e-05,
+      "loss": 0.1659,
+      "step": 12100
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 4.347695990424895e-05,
+      "loss": 0.1545,
+      "step": 12110
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.346698583682426e-05,
+      "loss": 0.0875,
+      "step": 12120
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 4.345701176939956e-05,
+      "loss": 0.1203,
+      "step": 12130
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 4.344703770197487e-05,
+      "loss": 0.1312,
+      "step": 12140
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.3437063634550175e-05,
+      "loss": 0.1163,
+      "step": 12150
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 4.342708956712548e-05,
+      "loss": 0.1216,
+      "step": 12160
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 4.341711549970078e-05,
+      "loss": 0.1751,
+      "step": 12170
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 4.340714143227608e-05,
+      "loss": 0.1181,
+      "step": 12180
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.339716736485139e-05,
+      "loss": 0.1294,
+      "step": 12190
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.338719329742669e-05,
+      "loss": 0.139,
+      "step": 12200
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 4.3377219230002e-05,
+      "loss": 0.1275,
+      "step": 12210
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 4.3367245162577305e-05,
+      "loss": 0.139,
+      "step": 12220
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 4.3357271095152605e-05,
+      "loss": 0.1283,
+      "step": 12230
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 4.334729702772791e-05,
+      "loss": 0.1268,
+      "step": 12240
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 4.333732296030321e-05,
+      "loss": 0.1665,
+      "step": 12250
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.3327348892878516e-05,
+      "loss": 0.1376,
+      "step": 12260
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.9692756223368468
+      },
+      "eval_f1": {
+        "f1": 0.9676560659554989
+      },
+      "eval_loss": 0.08475445210933685,
+      "eval_precision": {
+        "precision": 0.9695619317426214
+      },
+      "eval_recall": {
+        "recall": 0.9663476308224139
+      },
+      "eval_runtime": 161.2544,
+      "eval_samples_per_second": 110.608,
+      "eval_steps_per_second": 6.915,
+      "step": 12262
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 4.331737482545382e-05,
+      "loss": 0.0999,
+      "step": 12270
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.330740075802913e-05,
+      "loss": 0.129,
+      "step": 12280
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.329742669060443e-05,
+      "loss": 0.1235,
+      "step": 12290
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.3287452623179734e-05,
+      "loss": 0.1092,
+      "step": 12300
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 4.327747855575504e-05,
+      "loss": 0.0992,
+      "step": 12310
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 4.326750448833034e-05,
+      "loss": 0.1047,
+      "step": 12320
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 4.3257530420905646e-05,
+      "loss": 0.1063,
+      "step": 12330
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 4.324755635348095e-05,
+      "loss": 0.146,
+      "step": 12340
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 4.323758228605625e-05,
+      "loss": 0.1151,
+      "step": 12350
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 4.322760821863156e-05,
+      "loss": 0.1213,
+      "step": 12360
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 4.3217634151206864e-05,
+      "loss": 0.1031,
+      "step": 12370
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 4.320766008378217e-05,
+      "loss": 0.1494,
+      "step": 12380
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 4.3197686016357477e-05,
+      "loss": 0.1238,
+      "step": 12390
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 4.3187711948932776e-05,
+      "loss": 0.1452,
+      "step": 12400
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 4.317773788150808e-05,
+      "loss": 0.1014,
+      "step": 12410
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 4.316776381408338e-05,
+      "loss": 0.1197,
+      "step": 12420
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 4.315778974665869e-05,
+      "loss": 0.1303,
+      "step": 12430
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 4.3147815679233994e-05,
+      "loss": 0.137,
+      "step": 12440
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 4.31378416118093e-05,
+      "loss": 0.116,
+      "step": 12450
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 4.3127867544384606e-05,
+      "loss": 0.1482,
+      "step": 12460
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 4.3117893476959906e-05,
+      "loss": 0.1227,
+      "step": 12470
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 4.3107919409535205e-05,
+      "loss": 0.1016,
+      "step": 12480
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 4.309794534211051e-05,
+      "loss": 0.1303,
+      "step": 12490
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 4.308797127468582e-05,
+      "loss": 0.12,
+      "step": 12500
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 4.3077997207261124e-05,
+      "loss": 0.1111,
+      "step": 12510
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 4.306802313983643e-05,
+      "loss": 0.1238,
+      "step": 12520
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 4.305804907241173e-05,
+      "loss": 0.1352,
+      "step": 12530
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 4.3048075004987036e-05,
+      "loss": 0.122,
+      "step": 12540
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 4.303810093756234e-05,
+      "loss": 0.1629,
+      "step": 12550
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 4.302812687013764e-05,
+      "loss": 0.1072,
+      "step": 12560
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 4.301815280271295e-05,
+      "loss": 0.142,
+      "step": 12570
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 4.3008178735288254e-05,
+      "loss": 0.1332,
+      "step": 12580
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 4.299820466786356e-05,
+      "loss": 0.131,
+      "step": 12590
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 4.298823060043886e-05,
+      "loss": 0.1203,
+      "step": 12600
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 4.2978256533014165e-05,
+      "loss": 0.1271,
+      "step": 12610
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.296828246558947e-05,
+      "loss": 0.132,
+      "step": 12620
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 4.295830839816478e-05,
+      "loss": 0.1323,
+      "step": 12630
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 4.294833433074008e-05,
+      "loss": 0.1193,
+      "step": 12640
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 4.293836026331538e-05,
+      "loss": 0.1116,
+      "step": 12650
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 4.292838619589068e-05,
+      "loss": 0.1064,
+      "step": 12660
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 4.291841212846599e-05,
+      "loss": 0.1123,
+      "step": 12670
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 4.2908438061041295e-05,
+      "loss": 0.0981,
+      "step": 12680
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 4.28984639936166e-05,
+      "loss": 0.1274,
+      "step": 12690
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 4.288848992619191e-05,
+      "loss": 0.123,
+      "step": 12700
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 4.287851585876721e-05,
+      "loss": 0.1028,
+      "step": 12710
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 4.2868541791342506e-05,
+      "loss": 0.1031,
+      "step": 12720
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 4.285856772391781e-05,
+      "loss": 0.1385,
+      "step": 12730
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 4.284859365649312e-05,
+      "loss": 0.1075,
+      "step": 12740
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 4.2838619589068425e-05,
+      "loss": 0.1079,
+      "step": 12750
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 4.282864552164373e-05,
+      "loss": 0.093,
+      "step": 12760
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 4.281867145421903e-05,
+      "loss": 0.1047,
+      "step": 12770
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 4.280869738679434e-05,
+      "loss": 0.1422,
+      "step": 12780
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 4.279872331936964e-05,
+      "loss": 0.1274,
+      "step": 12790
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 4.278874925194494e-05,
+      "loss": 0.1199,
+      "step": 12800
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 4.277877518452025e-05,
+      "loss": 0.1427,
+      "step": 12810
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 4.2768801117095555e-05,
+      "loss": 0.1538,
+      "step": 12820
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 4.275882704967086e-05,
+      "loss": 0.1497,
+      "step": 12830
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 4.274885298224616e-05,
+      "loss": 0.104,
+      "step": 12840
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 4.2738878914821467e-05,
+      "loss": 0.1474,
+      "step": 12850
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 4.272890484739677e-05,
+      "loss": 0.1277,
+      "step": 12860
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 4.271893077997208e-05,
+      "loss": 0.1025,
+      "step": 12870
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 4.270895671254738e-05,
+      "loss": 0.1431,
+      "step": 12880
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 4.2698982645122685e-05,
+      "loss": 0.1329,
+      "step": 12890
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 4.2689008577697984e-05,
+      "loss": 0.1403,
+      "step": 12900
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 4.267903451027329e-05,
+      "loss": 0.1341,
+      "step": 12910
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 4.2669060442848596e-05,
+      "loss": 0.1234,
+      "step": 12920
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 4.26590863754239e-05,
+      "loss": 0.1007,
+      "step": 12930
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 4.264911230799921e-05,
+      "loss": 0.1352,
+      "step": 12940
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.263913824057451e-05,
+      "loss": 0.1197,
+      "step": 12950
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 4.262916417314981e-05,
+      "loss": 0.1119,
+      "step": 12960
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 4.2619190105725114e-05,
+      "loss": 0.1086,
+      "step": 12970
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 4.260921603830042e-05,
+      "loss": 0.0853,
+      "step": 12980
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 4.2599241970875726e-05,
+      "loss": 0.111,
+      "step": 12990
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 4.258926790345103e-05,
+      "loss": 0.1335,
+      "step": 13000
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 4.257929383602633e-05,
+      "loss": 0.1439,
+      "step": 13010
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 4.256931976860164e-05,
+      "loss": 0.098,
+      "step": 13020
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 4.2559345701176944e-05,
+      "loss": 0.1288,
+      "step": 13030
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 4.2549371633752244e-05,
+      "loss": 0.1038,
+      "step": 13040
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 4.253939756632755e-05,
+      "loss": 0.1233,
+      "step": 13050
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 4.2529423498902856e-05,
+      "loss": 0.1029,
+      "step": 13060
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 4.251944943147816e-05,
+      "loss": 0.1227,
+      "step": 13070
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 4.250947536405346e-05,
+      "loss": 0.1554,
+      "step": 13080
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 4.249950129662877e-05,
+      "loss": 0.1199,
+      "step": 13090
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 4.2489527229204074e-05,
+      "loss": 0.1324,
+      "step": 13100
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 4.247955316177937e-05,
+      "loss": 0.1506,
+      "step": 13110
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 4.246957909435468e-05,
+      "loss": 0.1075,
+      "step": 13120
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 4.2459605026929986e-05,
+      "loss": 0.1315,
+      "step": 13130
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 4.2449630959505285e-05,
+      "loss": 0.1372,
+      "step": 13140
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 4.243965689208059e-05,
+      "loss": 0.1101,
+      "step": 13150
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 4.24296828246559e-05,
+      "loss": 0.1091,
+      "step": 13160
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 4.2419708757231204e-05,
+      "loss": 0.1057,
+      "step": 13170
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 4.240973468980651e-05,
+      "loss": 0.1167,
+      "step": 13180
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 4.239976062238181e-05,
+      "loss": 0.1252,
+      "step": 13190
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 4.238978655495711e-05,
+      "loss": 0.1235,
+      "step": 13200
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 4.2379812487532415e-05,
+      "loss": 0.0959,
+      "step": 13210
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 4.236983842010772e-05,
+      "loss": 0.1364,
+      "step": 13220
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 4.235986435268303e-05,
+      "loss": 0.1285,
+      "step": 13230
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 4.2349890285258333e-05,
+      "loss": 0.1142,
+      "step": 13240
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.233991621783363e-05,
+      "loss": 0.1377,
+      "step": 13250
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.232994215040894e-05,
+      "loss": 0.1199,
+      "step": 13260
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.231996808298424e-05,
+      "loss": 0.1047,
+      "step": 13270
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 4.2309994015559545e-05,
+      "loss": 0.1015,
+      "step": 13280
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.230001994813485e-05,
+      "loss": 0.1395,
+      "step": 13290
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 4.229004588071016e-05,
+      "loss": 0.0903,
+      "step": 13300
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 4.228007181328546e-05,
+      "loss": 0.1238,
+      "step": 13310
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 4.227009774586076e-05,
+      "loss": 0.1444,
+      "step": 13320
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 4.226012367843607e-05,
+      "loss": 0.1685,
+      "step": 13330
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 4.2250149611011375e-05,
+      "loss": 0.1144,
+      "step": 13340
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 4.2240175543586674e-05,
+      "loss": 0.116,
+      "step": 13350
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 4.223020147616198e-05,
+      "loss": 0.103,
+      "step": 13360
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 4.222022740873729e-05,
+      "loss": 0.1681,
+      "step": 13370
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9743776631531733
+      },
+      "eval_f1": {
+        "f1": 0.9731609997444961
+      },
+      "eval_loss": 0.07890967279672623,
+      "eval_precision": {
+        "precision": 0.973142077383281
+      },
+      "eval_recall": {
+        "recall": 0.9733127731609976
+      },
+      "eval_runtime": 161.9943,
+      "eval_samples_per_second": 110.103,
+      "eval_steps_per_second": 6.883,
+      "step": 13377
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 4.2210253341312586e-05,
+      "loss": 0.1354,
+      "step": 13380
+    },
+    {
+      "epoch": 12.01,
+      "learning_rate": 4.220027927388789e-05,
+      "loss": 0.0893,
+      "step": 13390
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 4.21903052064632e-05,
+      "loss": 0.1122,
+      "step": 13400
+    },
+    {
+      "epoch": 12.03,
+      "learning_rate": 4.2180331139038505e-05,
+      "loss": 0.0998,
+      "step": 13410
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 4.217035707161381e-05,
+      "loss": 0.0957,
+      "step": 13420
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 4.216038300418911e-05,
+      "loss": 0.1198,
+      "step": 13430
+    },
+    {
+      "epoch": 12.06,
+      "learning_rate": 4.215040893676441e-05,
+      "loss": 0.0966,
+      "step": 13440
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 4.2140434869339716e-05,
+      "loss": 0.1341,
+      "step": 13450
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 4.213046080191502e-05,
+      "loss": 0.1179,
+      "step": 13460
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 4.212048673449033e-05,
+      "loss": 0.0897,
+      "step": 13470
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 4.2110512667065635e-05,
+      "loss": 0.154,
+      "step": 13480
+    },
+    {
+      "epoch": 12.1,
+      "learning_rate": 4.2100538599640934e-05,
+      "loss": 0.1323,
+      "step": 13490
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 4.209056453221624e-05,
+      "loss": 0.0912,
+      "step": 13500
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 4.208059046479154e-05,
+      "loss": 0.1077,
+      "step": 13510
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 4.2070616397366846e-05,
+      "loss": 0.109,
+      "step": 13520
+    },
+    {
+      "epoch": 12.14,
+      "learning_rate": 4.206064232994215e-05,
+      "loss": 0.0932,
+      "step": 13530
+    },
+    {
+      "epoch": 12.15,
+      "learning_rate": 4.205066826251746e-05,
+      "loss": 0.0827,
+      "step": 13540
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 4.2040694195092764e-05,
+      "loss": 0.1725,
+      "step": 13550
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 4.2030720127668064e-05,
+      "loss": 0.0932,
+      "step": 13560
+    },
+    {
+      "epoch": 12.17,
+      "learning_rate": 4.202074606024337e-05,
+      "loss": 0.1352,
+      "step": 13570
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 4.2010771992818676e-05,
+      "loss": 0.1148,
+      "step": 13580
+    },
+    {
+      "epoch": 12.19,
+      "learning_rate": 4.2000797925393976e-05,
+      "loss": 0.1041,
+      "step": 13590
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 4.199082385796928e-05,
+      "loss": 0.0899,
+      "step": 13600
+    },
+    {
+      "epoch": 12.21,
+      "learning_rate": 4.198084979054459e-05,
+      "loss": 0.115,
+      "step": 13610
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 4.197087572311989e-05,
+      "loss": 0.1415,
+      "step": 13620
+    },
+    {
+      "epoch": 12.23,
+      "learning_rate": 4.1960901655695194e-05,
+      "loss": 0.0972,
+      "step": 13630
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 4.19509275882705e-05,
+      "loss": 0.1286,
+      "step": 13640
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 4.1940953520845806e-05,
+      "loss": 0.1106,
+      "step": 13650
+    },
+    {
+      "epoch": 12.25,
+      "learning_rate": 4.193097945342111e-05,
+      "loss": 0.0992,
+      "step": 13660
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 4.192100538599641e-05,
+      "loss": 0.1457,
+      "step": 13670
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 4.191103131857171e-05,
+      "loss": 0.1567,
+      "step": 13680
+    },
+    {
+      "epoch": 12.28,
+      "learning_rate": 4.190105725114702e-05,
+      "loss": 0.1121,
+      "step": 13690
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 4.1891083183722323e-05,
+      "loss": 0.1257,
+      "step": 13700
+    },
+    {
+      "epoch": 12.3,
+      "learning_rate": 4.188110911629763e-05,
+      "loss": 0.1026,
+      "step": 13710
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 4.1871135048872936e-05,
+      "loss": 0.0881,
+      "step": 13720
+    },
+    {
+      "epoch": 12.32,
+      "learning_rate": 4.186116098144824e-05,
+      "loss": 0.1186,
+      "step": 13730
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 4.185118691402354e-05,
+      "loss": 0.1154,
+      "step": 13740
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 4.184121284659884e-05,
+      "loss": 0.1001,
+      "step": 13750
+    },
+    {
+      "epoch": 12.34,
+      "learning_rate": 4.183123877917415e-05,
+      "loss": 0.0714,
+      "step": 13760
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 4.182126471174945e-05,
+      "loss": 0.1276,
+      "step": 13770
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 4.181129064432476e-05,
+      "loss": 0.1192,
+      "step": 13780
+    },
+    {
+      "epoch": 12.37,
+      "learning_rate": 4.1801316576900066e-05,
+      "loss": 0.1328,
+      "step": 13790
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 4.1791342509475365e-05,
+      "loss": 0.08,
+      "step": 13800
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 4.178136844205067e-05,
+      "loss": 0.1459,
+      "step": 13810
+    },
+    {
+      "epoch": 12.4,
+      "learning_rate": 4.177139437462597e-05,
+      "loss": 0.1012,
+      "step": 13820
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 4.176142030720128e-05,
+      "loss": 0.0934,
+      "step": 13830
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 4.175144623977658e-05,
+      "loss": 0.1223,
+      "step": 13840
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 4.174147217235189e-05,
+      "loss": 0.0905,
+      "step": 13850
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 4.173149810492719e-05,
+      "loss": 0.1463,
+      "step": 13860
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 4.1721524037502495e-05,
+      "loss": 0.1266,
+      "step": 13870
+    },
+    {
+      "epoch": 12.45,
+      "learning_rate": 4.17115499700778e-05,
+      "loss": 0.1075,
+      "step": 13880
+    },
+    {
+      "epoch": 12.46,
+      "learning_rate": 4.170157590265311e-05,
+      "loss": 0.1003,
+      "step": 13890
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 4.1691601835228407e-05,
+      "loss": 0.1095,
+      "step": 13900
+    },
+    {
+      "epoch": 12.48,
+      "learning_rate": 4.168162776780371e-05,
+      "loss": 0.1053,
+      "step": 13910
+    },
+    {
+      "epoch": 12.49,
+      "learning_rate": 4.167165370037901e-05,
+      "loss": 0.1165,
+      "step": 13920
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 4.166167963295432e-05,
+      "loss": 0.1243,
+      "step": 13930
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 4.1651705565529625e-05,
+      "loss": 0.1329,
+      "step": 13940
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 4.164173149810493e-05,
+      "loss": 0.1065,
+      "step": 13950
+    },
+    {
+      "epoch": 12.52,
+      "learning_rate": 4.163175743068024e-05,
+      "loss": 0.1049,
+      "step": 13960
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 4.162178336325554e-05,
+      "loss": 0.1283,
+      "step": 13970
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 4.161180929583084e-05,
+      "loss": 0.1076,
+      "step": 13980
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 4.160183522840614e-05,
+      "loss": 0.1303,
+      "step": 13990
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 4.159186116098145e-05,
+      "loss": 0.1264,
+      "step": 14000
+    },
+    {
+      "epoch": 12.57,
+      "learning_rate": 4.1581887093556754e-05,
+      "loss": 0.1266,
+      "step": 14010
+    },
+    {
+      "epoch": 12.58,
+      "learning_rate": 4.157191302613206e-05,
+      "loss": 0.1347,
+      "step": 14020
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.156193895870737e-05,
+      "loss": 0.1209,
+      "step": 14030
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 4.1551964891282666e-05,
+      "loss": 0.1074,
+      "step": 14040
+    },
+    {
+      "epoch": 12.6,
+      "learning_rate": 4.154199082385797e-05,
+      "loss": 0.1025,
+      "step": 14050
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 4.153201675643327e-05,
+      "loss": 0.1173,
+      "step": 14060
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 4.152204268900858e-05,
+      "loss": 0.122,
+      "step": 14070
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 4.1512068621583884e-05,
+      "loss": 0.1153,
+      "step": 14080
+    },
+    {
+      "epoch": 12.64,
+      "learning_rate": 4.150209455415919e-05,
+      "loss": 0.1121,
+      "step": 14090
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 4.149212048673449e-05,
+      "loss": 0.1113,
+      "step": 14100
+    },
+    {
+      "epoch": 12.66,
+      "learning_rate": 4.1482146419309796e-05,
+      "loss": 0.0897,
+      "step": 14110
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 4.14721723518851e-05,
+      "loss": 0.0956,
+      "step": 14120
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.146219828446041e-05,
+      "loss": 0.0917,
+      "step": 14130
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 4.145222421703571e-05,
+      "loss": 0.1219,
+      "step": 14140
+    },
+    {
+      "epoch": 12.69,
+      "learning_rate": 4.1442250149611014e-05,
+      "loss": 0.1079,
+      "step": 14150
+    },
+    {
+      "epoch": 12.7,
+      "learning_rate": 4.143227608218631e-05,
+      "loss": 0.0977,
+      "step": 14160
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 4.142230201476162e-05,
+      "loss": 0.1064,
+      "step": 14170
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 4.1412327947336926e-05,
+      "loss": 0.1123,
+      "step": 14180
+    },
+    {
+      "epoch": 12.73,
+      "learning_rate": 4.140235387991223e-05,
+      "loss": 0.1227,
+      "step": 14190
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 4.139237981248754e-05,
+      "loss": 0.1035,
+      "step": 14200
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 4.1382405745062844e-05,
+      "loss": 0.1254,
+      "step": 14210
+    },
+    {
+      "epoch": 12.76,
+      "learning_rate": 4.1372431677638144e-05,
+      "loss": 0.1331,
+      "step": 14220
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 4.136245761021344e-05,
+      "loss": 0.1123,
+      "step": 14230
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 4.135248354278875e-05,
+      "loss": 0.0924,
+      "step": 14240
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 4.1342509475364056e-05,
+      "loss": 0.1135,
+      "step": 14250
+    },
+    {
+      "epoch": 12.79,
+      "learning_rate": 4.133253540793936e-05,
+      "loss": 0.1399,
+      "step": 14260
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 4.132256134051467e-05,
+      "loss": 0.1022,
+      "step": 14270
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 4.131258727308997e-05,
+      "loss": 0.1232,
+      "step": 14280
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 4.1302613205665274e-05,
+      "loss": 0.1135,
+      "step": 14290
+    },
+    {
+      "epoch": 12.83,
+      "learning_rate": 4.129263913824057e-05,
+      "loss": 0.0942,
+      "step": 14300
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 4.128266507081588e-05,
+      "loss": 0.1004,
+      "step": 14310
+    },
+    {
+      "epoch": 12.85,
+      "learning_rate": 4.1272691003391185e-05,
+      "loss": 0.1098,
+      "step": 14320
+    },
+    {
+      "epoch": 12.85,
+      "learning_rate": 4.126271693596649e-05,
+      "loss": 0.1168,
+      "step": 14330
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 4.125274286854179e-05,
+      "loss": 0.1083,
+      "step": 14340
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 4.12427688011171e-05,
+      "loss": 0.1093,
+      "step": 14350
+    },
+    {
+      "epoch": 12.88,
+      "learning_rate": 4.12327947336924e-05,
+      "loss": 0.1071,
+      "step": 14360
+    },
+    {
+      "epoch": 12.89,
+      "learning_rate": 4.122282066626771e-05,
+      "loss": 0.1147,
+      "step": 14370
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 4.121284659884301e-05,
+      "loss": 0.1371,
+      "step": 14380
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 4.1202872531418315e-05,
+      "loss": 0.11,
+      "step": 14390
+    },
+    {
+      "epoch": 12.92,
+      "learning_rate": 4.1192898463993615e-05,
+      "loss": 0.1167,
+      "step": 14400
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 4.118292439656892e-05,
+      "loss": 0.0954,
+      "step": 14410
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 4.117295032914423e-05,
+      "loss": 0.1242,
+      "step": 14420
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 4.116297626171953e-05,
+      "loss": 0.0974,
+      "step": 14430
+    },
+    {
+      "epoch": 12.95,
+      "learning_rate": 4.115300219429484e-05,
+      "loss": 0.0991,
+      "step": 14440
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 4.1143028126870145e-05,
+      "loss": 0.0988,
+      "step": 14450
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 4.1133054059445445e-05,
+      "loss": 0.0949,
+      "step": 14460
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 4.1123079992020744e-05,
+      "loss": 0.1157,
+      "step": 14470
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 4.111310592459605e-05,
+      "loss": 0.1037,
+      "step": 14480
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 4.110313185717136e-05,
+      "loss": 0.0998,
+      "step": 14490
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9749943933617403
+      },
+      "eval_f1": {
+        "f1": 0.9737724540493713
+      },
+      "eval_loss": 0.07498600333929062,
+      "eval_precision": {
+        "precision": 0.9739505389554447
+      },
+      "eval_recall": {
+        "recall": 0.9737043141281871
+      },
+      "eval_runtime": 172.5529,
+      "eval_samples_per_second": 103.365,
+      "eval_steps_per_second": 6.462,
+      "step": 14491
+    },
+    {
+      "epoch": 13.01,
+      "learning_rate": 4.109315778974666e-05,
+      "loss": 0.1154,
+      "step": 14500
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 4.108318372232197e-05,
+      "loss": 0.0994,
+      "step": 14510
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 4.107320965489727e-05,
+      "loss": 0.0747,
+      "step": 14520
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 4.1063235587472575e-05,
+      "loss": 0.0722,
+      "step": 14530
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 4.1053261520047874e-05,
+      "loss": 0.1177,
+      "step": 14540
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 4.104328745262318e-05,
+      "loss": 0.0982,
+      "step": 14550
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 4.1033313385198486e-05,
+      "loss": 0.127,
+      "step": 14560
+    },
+    {
+      "epoch": 13.07,
+      "learning_rate": 4.102333931777379e-05,
+      "loss": 0.1065,
+      "step": 14570
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 4.101336525034909e-05,
+      "loss": 0.1178,
+      "step": 14580
+    },
+    {
+      "epoch": 13.09,
+      "learning_rate": 4.10033911829244e-05,
+      "loss": 0.0873,
+      "step": 14590
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 4.0993417115499704e-05,
+      "loss": 0.1118,
+      "step": 14600
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 4.0983443048075004e-05,
+      "loss": 0.1068,
+      "step": 14610
+    },
+    {
+      "epoch": 13.12,
+      "learning_rate": 4.097346898065031e-05,
+      "loss": 0.1106,
+      "step": 14620
+    },
+    {
+      "epoch": 13.12,
+      "learning_rate": 4.0963494913225616e-05,
+      "loss": 0.1089,
+      "step": 14630
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 4.0953520845800916e-05,
+      "loss": 0.1607,
+      "step": 14640
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 4.094354677837622e-05,
+      "loss": 0.1101,
+      "step": 14650
+    },
+    {
+      "epoch": 13.15,
+      "learning_rate": 4.093357271095153e-05,
+      "loss": 0.1129,
+      "step": 14660
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 4.0923598643526834e-05,
+      "loss": 0.1169,
+      "step": 14670
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 4.091362457610214e-05,
+      "loss": 0.1141,
+      "step": 14680
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 4.090365050867744e-05,
+      "loss": 0.1041,
+      "step": 14690
+    },
+    {
+      "epoch": 13.19,
+      "learning_rate": 4.0893676441252746e-05,
+      "loss": 0.1018,
+      "step": 14700
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 4.0883702373828045e-05,
+      "loss": 0.1089,
+      "step": 14710
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 4.087372830640335e-05,
+      "loss": 0.1105,
+      "step": 14720
+    },
+    {
+      "epoch": 13.21,
+      "learning_rate": 4.086375423897866e-05,
+      "loss": 0.1029,
+      "step": 14730
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 4.0853780171553964e-05,
+      "loss": 0.1364,
+      "step": 14740
+    },
+    {
+      "epoch": 13.23,
+      "learning_rate": 4.084380610412927e-05,
+      "loss": 0.1041,
+      "step": 14750
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 4.083383203670457e-05,
+      "loss": 0.1103,
+      "step": 14760
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 4.082385796927987e-05,
+      "loss": 0.1155,
+      "step": 14770
+    },
+    {
+      "epoch": 13.26,
+      "learning_rate": 4.0813883901855175e-05,
+      "loss": 0.1258,
+      "step": 14780
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 4.080390983443048e-05,
+      "loss": 0.0947,
+      "step": 14790
+    },
+    {
+      "epoch": 13.28,
+      "learning_rate": 4.079393576700579e-05,
+      "loss": 0.0969,
+      "step": 14800
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 4.0783961699581094e-05,
+      "loss": 0.1169,
+      "step": 14810
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 4.077398763215639e-05,
+      "loss": 0.0877,
+      "step": 14820
+    },
+    {
+      "epoch": 13.3,
+      "learning_rate": 4.07640135647317e-05,
+      "loss": 0.0979,
+      "step": 14830
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 4.0754039497307006e-05,
+      "loss": 0.0894,
+      "step": 14840
+    },
+    {
+      "epoch": 13.32,
+      "learning_rate": 4.0744065429882305e-05,
+      "loss": 0.1537,
+      "step": 14850
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 4.073409136245761e-05,
+      "loss": 0.0964,
+      "step": 14860
+    },
+    {
+      "epoch": 13.34,
+      "learning_rate": 4.072411729503292e-05,
+      "loss": 0.1393,
+      "step": 14870
+    },
+    {
+      "epoch": 13.35,
+      "learning_rate": 4.071414322760822e-05,
+      "loss": 0.1681,
+      "step": 14880
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 4.070416916018352e-05,
+      "loss": 0.1072,
+      "step": 14890
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 4.069419509275883e-05,
+      "loss": 0.0971,
+      "step": 14900
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 4.0684221025334135e-05,
+      "loss": 0.0972,
+      "step": 14910
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 4.067424695790944e-05,
+      "loss": 0.1325,
+      "step": 14920
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 4.066427289048474e-05,
+      "loss": 0.1205,
+      "step": 14930
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 4.065429882306005e-05,
+      "loss": 0.0846,
+      "step": 14940
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 4.064432475563535e-05,
+      "loss": 0.1352,
+      "step": 14950
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 4.063435068821065e-05,
+      "loss": 0.1269,
+      "step": 14960
+    },
+    {
+      "epoch": 13.43,
+      "learning_rate": 4.062437662078596e-05,
+      "loss": 0.1143,
+      "step": 14970
+    },
+    {
+      "epoch": 13.44,
+      "learning_rate": 4.0614402553361265e-05,
+      "loss": 0.0854,
+      "step": 14980
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 4.060442848593657e-05,
+      "loss": 0.0991,
+      "step": 14990
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 4.059445441851187e-05,
+      "loss": 0.1022,
+      "step": 15000
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 4.058448035108717e-05,
+      "loss": 0.1268,
+      "step": 15010
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 4.0574506283662476e-05,
+      "loss": 0.0857,
+      "step": 15020
+    },
+    {
+      "epoch": 13.48,
+      "learning_rate": 4.056453221623778e-05,
+      "loss": 0.1497,
+      "step": 15030
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 4.055455814881309e-05,
+      "loss": 0.0904,
+      "step": 15040
+    },
+    {
+      "epoch": 13.5,
+      "learning_rate": 4.0544584081388395e-05,
+      "loss": 0.128,
+      "step": 15050
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 4.0534610013963694e-05,
+      "loss": 0.1283,
+      "step": 15060
+    },
+    {
+      "epoch": 13.52,
+      "learning_rate": 4.0524635946539e-05,
+      "loss": 0.1183,
+      "step": 15070
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 4.051466187911431e-05,
+      "loss": 0.1405,
+      "step": 15080
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 4.0504687811689606e-05,
+      "loss": 0.1043,
+      "step": 15090
+    },
+    {
+      "epoch": 13.55,
+      "learning_rate": 4.049471374426491e-05,
+      "loss": 0.0991,
+      "step": 15100
+    },
+    {
+      "epoch": 13.55,
+      "learning_rate": 4.048473967684022e-05,
+      "loss": 0.1278,
+      "step": 15110
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 4.0474765609415525e-05,
+      "loss": 0.1198,
+      "step": 15120
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 4.0464791541990824e-05,
+      "loss": 0.0714,
+      "step": 15130
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 4.045481747456613e-05,
+      "loss": 0.1306,
+      "step": 15140
+    },
+    {
+      "epoch": 13.59,
+      "learning_rate": 4.0444843407141437e-05,
+      "loss": 0.0716,
+      "step": 15150
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 4.043486933971674e-05,
+      "loss": 0.1188,
+      "step": 15160
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 4.042489527229204e-05,
+      "loss": 0.078,
+      "step": 15170
+    },
+    {
+      "epoch": 13.62,
+      "learning_rate": 4.041492120486735e-05,
+      "loss": 0.1037,
+      "step": 15180
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 4.040494713744265e-05,
+      "loss": 0.1074,
+      "step": 15190
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 4.0394973070017954e-05,
+      "loss": 0.1003,
+      "step": 15200
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 4.038499900259326e-05,
+      "loss": 0.1538,
+      "step": 15210
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 4.0375024935168566e-05,
+      "loss": 0.1044,
+      "step": 15220
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 4.036505086774387e-05,
+      "loss": 0.1202,
+      "step": 15230
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 4.035507680031917e-05,
+      "loss": 0.1156,
+      "step": 15240
+    },
+    {
+      "epoch": 13.68,
+      "learning_rate": 4.034510273289447e-05,
+      "loss": 0.1022,
+      "step": 15250
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 4.033512866546978e-05,
+      "loss": 0.1139,
+      "step": 15260
+    },
+    {
+      "epoch": 13.7,
+      "learning_rate": 4.0325154598045084e-05,
+      "loss": 0.1309,
+      "step": 15270
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 4.031518053062039e-05,
+      "loss": 0.1251,
+      "step": 15280
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 4.0305206463195696e-05,
+      "loss": 0.0879,
+      "step": 15290
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 4.0295232395770996e-05,
+      "loss": 0.0786,
+      "step": 15300
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 4.02852583283463e-05,
+      "loss": 0.089,
+      "step": 15310
+    },
+    {
+      "epoch": 13.74,
+      "learning_rate": 4.02752842609216e-05,
+      "loss": 0.1149,
+      "step": 15320
+    },
+    {
+      "epoch": 13.75,
+      "learning_rate": 4.026531019349691e-05,
+      "loss": 0.0929,
+      "step": 15330
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 4.0255336126072214e-05,
+      "loss": 0.0804,
+      "step": 15340
+    },
+    {
+      "epoch": 13.77,
+      "learning_rate": 4.024536205864752e-05,
+      "loss": 0.1287,
+      "step": 15350
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 4.0235387991222826e-05,
+      "loss": 0.0876,
+      "step": 15360
+    },
+    {
+      "epoch": 13.79,
+      "learning_rate": 4.0225413923798125e-05,
+      "loss": 0.131,
+      "step": 15370
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 4.021543985637343e-05,
+      "loss": 0.105,
+      "step": 15380
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 4.020546578894874e-05,
+      "loss": 0.0976,
+      "step": 15390
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 4.019549172152404e-05,
+      "loss": 0.0979,
+      "step": 15400
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 4.018551765409934e-05,
+      "loss": 0.0895,
+      "step": 15410
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 4.017554358667465e-05,
+      "loss": 0.1029,
+      "step": 15420
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 4.016556951924995e-05,
+      "loss": 0.0611,
+      "step": 15430
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 4.0155595451825255e-05,
+      "loss": 0.1051,
+      "step": 15440
+    },
+    {
+      "epoch": 13.86,
+      "learning_rate": 4.014562138440056e-05,
+      "loss": 0.1059,
+      "step": 15450
+    },
+    {
+      "epoch": 13.87,
+      "learning_rate": 4.013564731697587e-05,
+      "loss": 0.1386,
+      "step": 15460
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 4.0125673249551174e-05,
+      "loss": 0.1037,
+      "step": 15470
+    },
+    {
+      "epoch": 13.89,
+      "learning_rate": 4.011569918212647e-05,
+      "loss": 0.115,
+      "step": 15480
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 4.010572511470177e-05,
+      "loss": 0.0987,
+      "step": 15490
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 4.009575104727708e-05,
+      "loss": 0.1062,
+      "step": 15500
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 4.0085776979852385e-05,
+      "loss": 0.0999,
+      "step": 15510
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 4.007580291242769e-05,
+      "loss": 0.101,
+      "step": 15520
+    },
+    {
+      "epoch": 13.93,
+      "learning_rate": 4.0065828845003e-05,
+      "loss": 0.1343,
+      "step": 15530
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 4.00558547775783e-05,
+      "loss": 0.1048,
+      "step": 15540
+    },
+    {
+      "epoch": 13.95,
+      "learning_rate": 4.00458807101536e-05,
+      "loss": 0.0906,
+      "step": 15550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 4.00359066427289e-05,
+      "loss": 0.1213,
+      "step": 15560
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 4.002593257530421e-05,
+      "loss": 0.098,
+      "step": 15570
+    },
+    {
+      "epoch": 13.98,
+      "learning_rate": 4.0015958507879515e-05,
+      "loss": 0.0989,
+      "step": 15580
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 4.000598444045482e-05,
+      "loss": 0.1009,
+      "step": 15590
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 3.999601037303013e-05,
+      "loss": 0.0968,
+      "step": 15600
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9737609329446064
+      },
+      "eval_f1": {
+        "f1": 0.9722837146803432
+      },
+      "eval_loss": 0.08372773230075836,
+      "eval_precision": {
+        "precision": 0.9729901044455692
+      },
+      "eval_recall": {
+        "recall": 0.9716272189580044
+      },
+      "eval_runtime": 170.2624,
+      "eval_samples_per_second": 104.756,
+      "eval_steps_per_second": 6.549,
+      "step": 15606
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.9986036305605427e-05,
+      "loss": 0.0796,
+      "step": 15610
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 3.997606223818073e-05,
+      "loss": 0.1053,
+      "step": 15620
+    },
+    {
+      "epoch": 14.02,
+      "learning_rate": 3.996608817075604e-05,
+      "loss": 0.0848,
+      "step": 15630
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 3.995611410333134e-05,
+      "loss": 0.1338,
+      "step": 15640
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 3.9946140035906645e-05,
+      "loss": 0.1082,
+      "step": 15650
+    },
+    {
+      "epoch": 14.05,
+      "learning_rate": 3.993616596848195e-05,
+      "loss": 0.1077,
+      "step": 15660
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 3.992619190105725e-05,
+      "loss": 0.1097,
+      "step": 15670
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 3.9916217833632556e-05,
+      "loss": 0.0874,
+      "step": 15680
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 3.990624376620786e-05,
+      "loss": 0.128,
+      "step": 15690
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 3.989626969878317e-05,
+      "loss": 0.0785,
+      "step": 15700
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 3.9886295631358475e-05,
+      "loss": 0.1114,
+      "step": 15710
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 3.9876321563933774e-05,
+      "loss": 0.1153,
+      "step": 15720
+    },
+    {
+      "epoch": 14.11,
+      "learning_rate": 3.9866347496509074e-05,
+      "loss": 0.1061,
+      "step": 15730
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 3.985637342908438e-05,
+      "loss": 0.0764,
+      "step": 15740
+    },
+    {
+      "epoch": 14.13,
+      "learning_rate": 3.9846399361659686e-05,
+      "loss": 0.0995,
+      "step": 15750
+    },
+    {
+      "epoch": 14.14,
+      "learning_rate": 3.983642529423499e-05,
+      "loss": 0.1197,
+      "step": 15760
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 3.98264512268103e-05,
+      "loss": 0.0706,
+      "step": 15770
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 3.98164771593856e-05,
+      "loss": 0.0785,
+      "step": 15780
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 3.9806503091960904e-05,
+      "loss": 0.1034,
+      "step": 15790
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 3.9796529024536204e-05,
+      "loss": 0.1011,
+      "step": 15800
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 3.978655495711151e-05,
+      "loss": 0.1235,
+      "step": 15810
+    },
+    {
+      "epoch": 14.19,
+      "learning_rate": 3.9776580889686816e-05,
+      "loss": 0.1249,
+      "step": 15820
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 3.976660682226212e-05,
+      "loss": 0.0985,
+      "step": 15830
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 3.975663275483743e-05,
+      "loss": 0.1329,
+      "step": 15840
+    },
+    {
+      "epoch": 14.22,
+      "learning_rate": 3.974665868741273e-05,
+      "loss": 0.0857,
+      "step": 15850
+    },
+    {
+      "epoch": 14.23,
+      "learning_rate": 3.9736684619988034e-05,
+      "loss": 0.1094,
+      "step": 15860
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 3.972671055256334e-05,
+      "loss": 0.1267,
+      "step": 15870
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 3.971673648513864e-05,
+      "loss": 0.1014,
+      "step": 15880
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 3.9706762417713946e-05,
+      "loss": 0.1012,
+      "step": 15890
+    },
+    {
+      "epoch": 14.26,
+      "learning_rate": 3.969678835028925e-05,
+      "loss": 0.09,
+      "step": 15900
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 3.968681428286455e-05,
+      "loss": 0.0867,
+      "step": 15910
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 3.967684021543986e-05,
+      "loss": 0.0931,
+      "step": 15920
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 3.9666866148015164e-05,
+      "loss": 0.0928,
+      "step": 15930
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 3.965689208059047e-05,
+      "loss": 0.0676,
+      "step": 15940
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 3.9646918013165776e-05,
+      "loss": 0.0823,
+      "step": 15950
+    },
+    {
+      "epoch": 14.32,
+      "learning_rate": 3.9636943945741075e-05,
+      "loss": 0.1044,
+      "step": 15960
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 3.9626969878316375e-05,
+      "loss": 0.09,
+      "step": 15970
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 3.961699581089168e-05,
+      "loss": 0.0941,
+      "step": 15980
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 3.960702174346699e-05,
+      "loss": 0.0897,
+      "step": 15990
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 3.9597047676042293e-05,
+      "loss": 0.0934,
+      "step": 16000
+    },
+    {
+      "epoch": 14.36,
+      "learning_rate": 3.95870736086176e-05,
+      "loss": 0.1458,
+      "step": 16010
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 3.95770995411929e-05,
+      "loss": 0.0936,
+      "step": 16020
+    },
+    {
+      "epoch": 14.38,
+      "learning_rate": 3.9567125473768205e-05,
+      "loss": 0.1156,
+      "step": 16030
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 3.9557151406343505e-05,
+      "loss": 0.1378,
+      "step": 16040
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 3.954717733891881e-05,
+      "loss": 0.0886,
+      "step": 16050
+    },
+    {
+      "epoch": 14.41,
+      "learning_rate": 3.953720327149412e-05,
+      "loss": 0.1047,
+      "step": 16060
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 3.952722920406942e-05,
+      "loss": 0.1054,
+      "step": 16070
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 3.951725513664473e-05,
+      "loss": 0.073,
+      "step": 16080
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 3.950728106922003e-05,
+      "loss": 0.1193,
+      "step": 16090
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 3.9497307001795335e-05,
+      "loss": 0.0978,
+      "step": 16100
+    },
+    {
+      "epoch": 14.45,
+      "learning_rate": 3.9487332934370634e-05,
+      "loss": 0.1323,
+      "step": 16110
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 3.947735886694594e-05,
+      "loss": 0.1283,
+      "step": 16120
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 3.946738479952125e-05,
+      "loss": 0.1021,
+      "step": 16130
+    },
+    {
+      "epoch": 14.48,
+      "learning_rate": 3.945741073209655e-05,
+      "loss": 0.116,
+      "step": 16140
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 3.944743666467185e-05,
+      "loss": 0.074,
+      "step": 16150
+    },
+    {
+      "epoch": 14.5,
+      "learning_rate": 3.943746259724716e-05,
+      "loss": 0.1102,
+      "step": 16160
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 3.9427488529822465e-05,
+      "loss": 0.0895,
+      "step": 16170
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 3.941751446239777e-05,
+      "loss": 0.1106,
+      "step": 16180
+    },
+    {
+      "epoch": 14.52,
+      "learning_rate": 3.940754039497307e-05,
+      "loss": 0.1006,
+      "step": 16190
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 3.939756632754838e-05,
+      "loss": 0.1123,
+      "step": 16200
+    },
+    {
+      "epoch": 14.54,
+      "learning_rate": 3.9387592260123676e-05,
+      "loss": 0.1302,
+      "step": 16210
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 3.937761819269898e-05,
+      "loss": 0.0982,
+      "step": 16220
+    },
+    {
+      "epoch": 14.56,
+      "learning_rate": 3.936764412527429e-05,
+      "loss": 0.099,
+      "step": 16230
+    },
+    {
+      "epoch": 14.57,
+      "learning_rate": 3.9357670057849595e-05,
+      "loss": 0.1191,
+      "step": 16240
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 3.93476959904249e-05,
+      "loss": 0.1029,
+      "step": 16250
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 3.933772192300021e-05,
+      "loss": 0.079,
+      "step": 16260
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.93277478555755e-05,
+      "loss": 0.106,
+      "step": 16270
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.9317773788150806e-05,
+      "loss": 0.1149,
+      "step": 16280
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 3.930779972072611e-05,
+      "loss": 0.0746,
+      "step": 16290
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 3.929782565330142e-05,
+      "loss": 0.1073,
+      "step": 16300
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 3.9287851585876724e-05,
+      "loss": 0.08,
+      "step": 16310
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 3.927787751845203e-05,
+      "loss": 0.1228,
+      "step": 16320
+    },
+    {
+      "epoch": 14.65,
+      "learning_rate": 3.926790345102733e-05,
+      "loss": 0.1137,
+      "step": 16330
+    },
+    {
+      "epoch": 14.66,
+      "learning_rate": 3.9257929383602636e-05,
+      "loss": 0.1034,
+      "step": 16340
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 3.9247955316177936e-05,
+      "loss": 0.0831,
+      "step": 16350
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 3.923798124875324e-05,
+      "loss": 0.1098,
+      "step": 16360
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 3.922800718132855e-05,
+      "loss": 0.1308,
+      "step": 16370
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 3.9218033113903854e-05,
+      "loss": 0.1113,
+      "step": 16380
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 3.9208059046479154e-05,
+      "loss": 0.1266,
+      "step": 16390
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 3.919808497905446e-05,
+      "loss": 0.1002,
+      "step": 16400
+    },
+    {
+      "epoch": 14.72,
+      "learning_rate": 3.9188110911629766e-05,
+      "loss": 0.0909,
+      "step": 16410
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 3.917813684420507e-05,
+      "loss": 0.1446,
+      "step": 16420
+    },
+    {
+      "epoch": 14.74,
+      "learning_rate": 3.916816277678037e-05,
+      "loss": 0.0919,
+      "step": 16430
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 3.915818870935568e-05,
+      "loss": 0.1184,
+      "step": 16440
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 3.914821464193098e-05,
+      "loss": 0.1161,
+      "step": 16450
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 3.9138240574506283e-05,
+      "loss": 0.091,
+      "step": 16460
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 3.912826650708159e-05,
+      "loss": 0.0807,
+      "step": 16470
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 3.9118292439656896e-05,
+      "loss": 0.0796,
+      "step": 16480
+    },
+    {
+      "epoch": 14.79,
+      "learning_rate": 3.91083183722322e-05,
+      "loss": 0.1246,
+      "step": 16490
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 3.909834430480751e-05,
+      "loss": 0.1058,
+      "step": 16500
+    },
+    {
+      "epoch": 14.81,
+      "learning_rate": 3.908837023738281e-05,
+      "loss": 0.0909,
+      "step": 16510
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 3.907839616995811e-05,
+      "loss": 0.081,
+      "step": 16520
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 3.906842210253341e-05,
+      "loss": 0.1071,
+      "step": 16530
+    },
+    {
+      "epoch": 14.84,
+      "learning_rate": 3.905844803510872e-05,
+      "loss": 0.101,
+      "step": 16540
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 3.9048473967684026e-05,
+      "loss": 0.12,
+      "step": 16550
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 3.903849990025933e-05,
+      "loss": 0.1199,
+      "step": 16560
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 3.902852583283463e-05,
+      "loss": 0.0715,
+      "step": 16570
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 3.901855176540994e-05,
+      "loss": 0.1047,
+      "step": 16580
+    },
+    {
+      "epoch": 14.88,
+      "learning_rate": 3.900857769798524e-05,
+      "loss": 0.1006,
+      "step": 16590
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 3.899860363056054e-05,
+      "loss": 0.1045,
+      "step": 16600
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 3.898862956313585e-05,
+      "loss": 0.0681,
+      "step": 16610
+    },
+    {
+      "epoch": 14.91,
+      "learning_rate": 3.8978655495711155e-05,
+      "loss": 0.1028,
+      "step": 16620
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 3.8968681428286455e-05,
+      "loss": 0.1052,
+      "step": 16630
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 3.895870736086176e-05,
+      "loss": 0.0892,
+      "step": 16640
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 3.894873329343707e-05,
+      "loss": 0.1078,
+      "step": 16650
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 3.893875922601237e-05,
+      "loss": 0.1642,
+      "step": 16660
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 3.892878515858767e-05,
+      "loss": 0.1068,
+      "step": 16670
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 3.891881109116298e-05,
+      "loss": 0.0962,
+      "step": 16680
+    },
+    {
+      "epoch": 14.97,
+      "learning_rate": 3.890883702373828e-05,
+      "loss": 0.0821,
+      "step": 16690
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 3.8898862956313585e-05,
+      "loss": 0.11,
+      "step": 16700
+    },
+    {
+      "epoch": 14.99,
+      "learning_rate": 3.888888888888889e-05,
+      "loss": 0.0895,
+      "step": 16710
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 3.88789148214642e-05,
+      "loss": 0.0841,
+      "step": 16720
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.9758914554832923
+      },
+      "eval_f1": {
+        "f1": 0.974522946671967
+      },
+      "eval_loss": 0.06894122064113617,
+      "eval_precision": {
+        "precision": 0.9749603470712005
+      },
+      "eval_recall": {
+        "recall": 0.974152667413215
+      },
+      "eval_runtime": 194.5243,
+      "eval_samples_per_second": 91.69,
+      "eval_steps_per_second": 5.732,
+      "step": 16721
+    },
+    {
+      "epoch": 15.01,
+      "learning_rate": 3.88689407540395e-05,
+      "loss": 0.0849,
+      "step": 16730
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 3.885896668661481e-05,
+      "loss": 0.1136,
+      "step": 16740
+    },
+    {
+      "epoch": 15.03,
+      "learning_rate": 3.884899261919011e-05,
+      "loss": 0.0709,
+      "step": 16750
+    },
+    {
+      "epoch": 15.03,
+      "learning_rate": 3.883901855176541e-05,
+      "loss": 0.097,
+      "step": 16760
+    },
+    {
+      "epoch": 15.04,
+      "learning_rate": 3.8829044484340714e-05,
+      "loss": 0.0982,
+      "step": 16770
+    },
+    {
+      "epoch": 15.05,
+      "learning_rate": 3.881907041691602e-05,
+      "loss": 0.1163,
+      "step": 16780
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 3.880909634949133e-05,
+      "loss": 0.1248,
+      "step": 16790
+    },
+    {
+      "epoch": 15.07,
+      "learning_rate": 3.879912228206663e-05,
+      "loss": 0.0843,
+      "step": 16800
+    },
+    {
+      "epoch": 15.08,
+      "learning_rate": 3.878914821464193e-05,
+      "loss": 0.0972,
+      "step": 16810
+    },
+    {
+      "epoch": 15.09,
+      "learning_rate": 3.877917414721723e-05,
+      "loss": 0.1123,
+      "step": 16820
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 3.876920007979254e-05,
+      "loss": 0.1049,
+      "step": 16830
+    },
+    {
+      "epoch": 15.11,
+      "learning_rate": 3.8759226012367844e-05,
+      "loss": 0.0983,
+      "step": 16840
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 3.874925194494315e-05,
+      "loss": 0.0821,
+      "step": 16850
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 3.8739277877518457e-05,
+      "loss": 0.0825,
+      "step": 16860
+    },
+    {
+      "epoch": 15.13,
+      "learning_rate": 3.8729303810093756e-05,
+      "loss": 0.098,
+      "step": 16870
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 3.871932974266906e-05,
+      "loss": 0.1046,
+      "step": 16880
+    },
+    {
+      "epoch": 15.15,
+      "learning_rate": 3.870935567524437e-05,
+      "loss": 0.0904,
+      "step": 16890
+    },
+    {
+      "epoch": 15.16,
+      "learning_rate": 3.869938160781967e-05,
+      "loss": 0.094,
+      "step": 16900
+    },
+    {
+      "epoch": 15.17,
+      "learning_rate": 3.8689407540394974e-05,
+      "loss": 0.0614,
+      "step": 16910
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 3.867943347297028e-05,
+      "loss": 0.1057,
+      "step": 16920
+    },
+    {
+      "epoch": 15.19,
+      "learning_rate": 3.866945940554558e-05,
+      "loss": 0.0773,
+      "step": 16930
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 3.8659485338120886e-05,
+      "loss": 0.0887,
+      "step": 16940
+    },
+    {
+      "epoch": 15.21,
+      "learning_rate": 3.864951127069619e-05,
+      "loss": 0.1208,
+      "step": 16950
+    },
+    {
+      "epoch": 15.21,
+      "learning_rate": 3.86395372032715e-05,
+      "loss": 0.1346,
+      "step": 16960
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 3.8629563135846804e-05,
+      "loss": 0.0987,
+      "step": 16970
+    },
+    {
+      "epoch": 15.23,
+      "learning_rate": 3.8619589068422104e-05,
+      "loss": 0.1247,
+      "step": 16980
+    },
+    {
+      "epoch": 15.24,
+      "learning_rate": 3.860961500099741e-05,
+      "loss": 0.102,
+      "step": 16990
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 3.859964093357271e-05,
+      "loss": 0.093,
+      "step": 17000
+    },
+    {
+      "epoch": 15.26,
+      "learning_rate": 3.8589666866148016e-05,
+      "loss": 0.105,
+      "step": 17010
+    },
+    {
+      "epoch": 15.27,
+      "learning_rate": 3.857969279872332e-05,
+      "loss": 0.0857,
+      "step": 17020
+    },
+    {
+      "epoch": 15.28,
+      "learning_rate": 3.856971873129863e-05,
+      "loss": 0.0871,
+      "step": 17030
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 3.8559744663873934e-05,
+      "loss": 0.1097,
+      "step": 17040
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 3.8549770596449234e-05,
+      "loss": 0.0946,
+      "step": 17050
+    },
+    {
+      "epoch": 15.3,
+      "learning_rate": 3.853979652902453e-05,
+      "loss": 0.0956,
+      "step": 17060
+    },
+    {
+      "epoch": 15.31,
+      "learning_rate": 3.852982246159984e-05,
+      "loss": 0.0801,
+      "step": 17070
+    },
+    {
+      "epoch": 15.32,
+      "learning_rate": 3.8519848394175145e-05,
+      "loss": 0.0743,
+      "step": 17080
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 3.850987432675045e-05,
+      "loss": 0.0605,
+      "step": 17090
+    },
+    {
+      "epoch": 15.34,
+      "learning_rate": 3.849990025932576e-05,
+      "loss": 0.0699,
+      "step": 17100
+    },
+    {
+      "epoch": 15.35,
+      "learning_rate": 3.848992619190106e-05,
+      "loss": 0.1319,
+      "step": 17110
+    },
+    {
+      "epoch": 15.36,
+      "learning_rate": 3.847995212447636e-05,
+      "loss": 0.097,
+      "step": 17120
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 3.846997805705167e-05,
+      "loss": 0.1144,
+      "step": 17130
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 3.846000398962697e-05,
+      "loss": 0.1205,
+      "step": 17140
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 3.8450029922202275e-05,
+      "loss": 0.1062,
+      "step": 17150
+    },
+    {
+      "epoch": 15.39,
+      "learning_rate": 3.844005585477758e-05,
+      "loss": 0.0882,
+      "step": 17160
+    },
+    {
+      "epoch": 15.4,
+      "learning_rate": 3.843008178735288e-05,
+      "loss": 0.0999,
+      "step": 17170
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 3.842010771992819e-05,
+      "loss": 0.1211,
+      "step": 17180
+    },
+    {
+      "epoch": 15.42,
+      "learning_rate": 3.841013365250349e-05,
+      "loss": 0.0659,
+      "step": 17190
+    },
+    {
+      "epoch": 15.43,
+      "learning_rate": 3.84001595850788e-05,
+      "loss": 0.0965,
+      "step": 17200
+    },
+    {
+      "epoch": 15.44,
+      "learning_rate": 3.8390185517654105e-05,
+      "loss": 0.0931,
+      "step": 17210
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 3.8380211450229405e-05,
+      "loss": 0.117,
+      "step": 17220
+    },
+    {
+      "epoch": 15.46,
+      "learning_rate": 3.837023738280471e-05,
+      "loss": 0.0961,
+      "step": 17230
+    },
+    {
+      "epoch": 15.47,
+      "learning_rate": 3.836026331538001e-05,
+      "loss": 0.1071,
+      "step": 17240
+    },
+    {
+      "epoch": 15.47,
+      "learning_rate": 3.835028924795532e-05,
+      "loss": 0.0997,
+      "step": 17250
+    },
+    {
+      "epoch": 15.48,
+      "learning_rate": 3.834031518053062e-05,
+      "loss": 0.1114,
+      "step": 17260
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 3.833034111310593e-05,
+      "loss": 0.0851,
+      "step": 17270
+    },
+    {
+      "epoch": 15.5,
+      "learning_rate": 3.8320367045681235e-05,
+      "loss": 0.0697,
+      "step": 17280
+    },
+    {
+      "epoch": 15.51,
+      "learning_rate": 3.8310392978256535e-05,
+      "loss": 0.1278,
+      "step": 17290
+    },
+    {
+      "epoch": 15.52,
+      "learning_rate": 3.8300418910831834e-05,
+      "loss": 0.1256,
+      "step": 17300
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 3.829044484340714e-05,
+      "loss": 0.1316,
+      "step": 17310
+    },
+    {
+      "epoch": 15.54,
+      "learning_rate": 3.8280470775982446e-05,
+      "loss": 0.0981,
+      "step": 17320
+    },
+    {
+      "epoch": 15.55,
+      "learning_rate": 3.827049670855775e-05,
+      "loss": 0.1178,
+      "step": 17330
+    },
+    {
+      "epoch": 15.56,
+      "learning_rate": 3.826052264113306e-05,
+      "loss": 0.1048,
+      "step": 17340
+    },
+    {
+      "epoch": 15.56,
+      "learning_rate": 3.825054857370836e-05,
+      "loss": 0.1117,
+      "step": 17350
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 3.8240574506283664e-05,
+      "loss": 0.1023,
+      "step": 17360
+    },
+    {
+      "epoch": 15.58,
+      "learning_rate": 3.823060043885897e-05,
+      "loss": 0.0686,
+      "step": 17370
+    },
+    {
+      "epoch": 15.59,
+      "learning_rate": 3.822062637143427e-05,
+      "loss": 0.1038,
+      "step": 17380
+    },
+    {
+      "epoch": 15.6,
+      "learning_rate": 3.8210652304009576e-05,
+      "loss": 0.0829,
+      "step": 17390
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 3.820067823658488e-05,
+      "loss": 0.0846,
+      "step": 17400
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 3.819070416916018e-05,
+      "loss": 0.0761,
+      "step": 17410
+    },
+    {
+      "epoch": 15.63,
+      "learning_rate": 3.818073010173549e-05,
+      "loss": 0.0772,
+      "step": 17420
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 3.8170756034310794e-05,
+      "loss": 0.0725,
+      "step": 17430
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 3.81607819668861e-05,
+      "loss": 0.1249,
+      "step": 17440
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 3.815080789946141e-05,
+      "loss": 0.1013,
+      "step": 17450
+    },
+    {
+      "epoch": 15.66,
+      "learning_rate": 3.8140833832036706e-05,
+      "loss": 0.1011,
+      "step": 17460
+    },
+    {
+      "epoch": 15.67,
+      "learning_rate": 3.813085976461201e-05,
+      "loss": 0.0907,
+      "step": 17470
+    },
+    {
+      "epoch": 15.68,
+      "learning_rate": 3.812088569718731e-05,
+      "loss": 0.0947,
+      "step": 17480
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 3.811091162976262e-05,
+      "loss": 0.1091,
+      "step": 17490
+    },
+    {
+      "epoch": 15.7,
+      "learning_rate": 3.8100937562337924e-05,
+      "loss": 0.0993,
+      "step": 17500
+    },
+    {
+      "epoch": 15.71,
+      "learning_rate": 3.809096349491323e-05,
+      "loss": 0.0751,
+      "step": 17510
+    },
+    {
+      "epoch": 15.72,
+      "learning_rate": 3.8080989427488536e-05,
+      "loss": 0.1132,
+      "step": 17520
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 3.8071015360063836e-05,
+      "loss": 0.1322,
+      "step": 17530
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 3.8061041292639135e-05,
+      "loss": 0.0828,
+      "step": 17540
+    },
+    {
+      "epoch": 15.74,
+      "learning_rate": 3.805106722521444e-05,
+      "loss": 0.1047,
+      "step": 17550
+    },
+    {
+      "epoch": 15.75,
+      "learning_rate": 3.804109315778975e-05,
+      "loss": 0.0997,
+      "step": 17560
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 3.8031119090365054e-05,
+      "loss": 0.1303,
+      "step": 17570
+    },
+    {
+      "epoch": 15.77,
+      "learning_rate": 3.802114502294036e-05,
+      "loss": 0.1073,
+      "step": 17580
+    },
+    {
+      "epoch": 15.78,
+      "learning_rate": 3.801117095551566e-05,
+      "loss": 0.099,
+      "step": 17590
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 3.8001196888090966e-05,
+      "loss": 0.0905,
+      "step": 17600
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 3.7991222820666265e-05,
+      "loss": 0.1211,
+      "step": 17610
+    },
+    {
+      "epoch": 15.81,
+      "learning_rate": 3.798124875324157e-05,
+      "loss": 0.1037,
+      "step": 17620
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 3.797127468581688e-05,
+      "loss": 0.0876,
+      "step": 17630
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 3.7961300618392184e-05,
+      "loss": 0.1117,
+      "step": 17640
+    },
+    {
+      "epoch": 15.83,
+      "learning_rate": 3.795132655096749e-05,
+      "loss": 0.1025,
+      "step": 17650
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 3.794135248354279e-05,
+      "loss": 0.0899,
+      "step": 17660
+    },
+    {
+      "epoch": 15.85,
+      "learning_rate": 3.7931378416118095e-05,
+      "loss": 0.0797,
+      "step": 17670
+    },
+    {
+      "epoch": 15.86,
+      "learning_rate": 3.79214043486934e-05,
+      "loss": 0.0716,
+      "step": 17680
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 3.79114302812687e-05,
+      "loss": 0.0742,
+      "step": 17690
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 3.790145621384401e-05,
+      "loss": 0.0849,
+      "step": 17700
+    },
+    {
+      "epoch": 15.89,
+      "learning_rate": 3.7891482146419313e-05,
+      "loss": 0.1001,
+      "step": 17710
+    },
+    {
+      "epoch": 15.9,
+      "learning_rate": 3.788150807899461e-05,
+      "loss": 0.11,
+      "step": 17720
+    },
+    {
+      "epoch": 15.9,
+      "learning_rate": 3.787153401156992e-05,
+      "loss": 0.0936,
+      "step": 17730
+    },
+    {
+      "epoch": 15.91,
+      "learning_rate": 3.7861559944145225e-05,
+      "loss": 0.1146,
+      "step": 17740
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 3.785158587672053e-05,
+      "loss": 0.0814,
+      "step": 17750
+    },
+    {
+      "epoch": 15.93,
+      "learning_rate": 3.784161180929584e-05,
+      "loss": 0.1027,
+      "step": 17760
+    },
+    {
+      "epoch": 15.94,
+      "learning_rate": 3.783163774187114e-05,
+      "loss": 0.1086,
+      "step": 17770
+    },
+    {
+      "epoch": 15.95,
+      "learning_rate": 3.7821663674446436e-05,
+      "loss": 0.1023,
+      "step": 17780
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 3.781168960702174e-05,
+      "loss": 0.0915,
+      "step": 17790
+    },
+    {
+      "epoch": 15.97,
+      "learning_rate": 3.780171553959705e-05,
+      "loss": 0.1068,
+      "step": 17800
+    },
+    {
+      "epoch": 15.98,
+      "learning_rate": 3.7791741472172355e-05,
+      "loss": 0.0943,
+      "step": 17810
+    },
+    {
+      "epoch": 15.99,
+      "learning_rate": 3.778176740474766e-05,
+      "loss": 0.1042,
+      "step": 17820
+    },
+    {
+      "epoch": 15.99,
+      "learning_rate": 3.777179333732296e-05,
+      "loss": 0.095,
+      "step": 17830
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": {
+        "accuracy": 0.9762278537788742
+      },
+      "eval_f1": {
+        "f1": 0.9749967013469043
+      },
+      "eval_loss": 0.07049217820167542,
+      "eval_precision": {
+        "precision": 0.9757537211246345
+      },
+      "eval_recall": {
+        "recall": 0.9743092871293193
+      },
+      "eval_runtime": 183.6354,
+      "eval_samples_per_second": 97.127,
+      "eval_steps_per_second": 6.072,
+      "step": 17836
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 3.776181926989827e-05,
+      "loss": 0.0805,
+      "step": 17840
+    },
+    {
+      "epoch": 16.01,
+      "learning_rate": 3.7751845202473566e-05,
+      "loss": 0.0908,
+      "step": 17850
+    },
+    {
+      "epoch": 16.02,
+      "learning_rate": 3.774187113504887e-05,
+      "loss": 0.0777,
+      "step": 17860
+    },
+    {
+      "epoch": 16.03,
+      "learning_rate": 3.773189706762418e-05,
+      "loss": 0.0984,
+      "step": 17870
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 3.7721923000199485e-05,
+      "loss": 0.1152,
+      "step": 17880
+    },
+    {
+      "epoch": 16.05,
+      "learning_rate": 3.771194893277479e-05,
+      "loss": 0.0914,
+      "step": 17890
+    },
+    {
+      "epoch": 16.06,
+      "learning_rate": 3.770197486535009e-05,
+      "loss": 0.0808,
+      "step": 17900
+    },
+    {
+      "epoch": 16.07,
+      "learning_rate": 3.7692000797925397e-05,
+      "loss": 0.0628,
+      "step": 17910
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 3.76820267305007e-05,
+      "loss": 0.0984,
+      "step": 17920
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 3.7672052663076e-05,
+      "loss": 0.0909,
+      "step": 17930
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 3.766207859565131e-05,
+      "loss": 0.0665,
+      "step": 17940
+    },
+    {
+      "epoch": 16.1,
+      "learning_rate": 3.7652104528226615e-05,
+      "loss": 0.1115,
+      "step": 17950
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 3.7642130460801914e-05,
+      "loss": 0.1008,
+      "step": 17960
+    },
+    {
+      "epoch": 16.12,
+      "learning_rate": 3.763215639337722e-05,
+      "loss": 0.0808,
+      "step": 17970
+    },
+    {
+      "epoch": 16.13,
+      "learning_rate": 3.7622182325952526e-05,
+      "loss": 0.0932,
+      "step": 17980
+    },
+    {
+      "epoch": 16.14,
+      "learning_rate": 3.761220825852783e-05,
+      "loss": 0.1051,
+      "step": 17990
+    },
+    {
+      "epoch": 16.15,
+      "learning_rate": 3.760223419110314e-05,
+      "loss": 0.0918,
+      "step": 18000
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 3.759226012367844e-05,
+      "loss": 0.1163,
+      "step": 18010
+    },
+    {
+      "epoch": 16.17,
+      "learning_rate": 3.758228605625374e-05,
+      "loss": 0.0997,
+      "step": 18020
+    },
+    {
+      "epoch": 16.17,
+      "learning_rate": 3.7572311988829044e-05,
+      "loss": 0.0933,
+      "step": 18030
+    },
+    {
+      "epoch": 16.18,
+      "learning_rate": 3.756233792140435e-05,
+      "loss": 0.0691,
+      "step": 18040
+    },
+    {
+      "epoch": 16.19,
+      "learning_rate": 3.7552363853979656e-05,
+      "loss": 0.0699,
+      "step": 18050
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 3.754238978655496e-05,
+      "loss": 0.1055,
+      "step": 18060
+    },
+    {
+      "epoch": 16.21,
+      "learning_rate": 3.753241571913026e-05,
+      "loss": 0.0703,
+      "step": 18070
+    },
+    {
+      "epoch": 16.22,
+      "learning_rate": 3.752244165170557e-05,
+      "loss": 0.0631,
+      "step": 18080
+    },
+    {
+      "epoch": 16.23,
+      "learning_rate": 3.751246758428087e-05,
+      "loss": 0.0824,
+      "step": 18090
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 3.7502493516856174e-05,
+      "loss": 0.0957,
+      "step": 18100
+    },
+    {
+      "epoch": 16.25,
+      "learning_rate": 3.749251944943148e-05,
+      "loss": 0.0816,
+      "step": 18110
+    },
+    {
+      "epoch": 16.25,
+      "learning_rate": 3.7482545382006786e-05,
+      "loss": 0.0799,
+      "step": 18120
+    },
+    {
+      "epoch": 16.26,
+      "learning_rate": 3.747257131458209e-05,
+      "loss": 0.0784,
+      "step": 18130
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 3.746259724715739e-05,
+      "loss": 0.1077,
+      "step": 18140
+    },
+    {
+      "epoch": 16.28,
+      "learning_rate": 3.74526231797327e-05,
+      "loss": 0.0862,
+      "step": 18150
+    },
+    {
+      "epoch": 16.29,
+      "learning_rate": 3.7442649112308004e-05,
+      "loss": 0.1011,
+      "step": 18160
+    },
+    {
+      "epoch": 16.3,
+      "learning_rate": 3.74326750448833e-05,
+      "loss": 0.1166,
+      "step": 18170
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 3.742270097745861e-05,
+      "loss": 0.1118,
+      "step": 18180
+    },
+    {
+      "epoch": 16.32,
+      "learning_rate": 3.7412726910033916e-05,
+      "loss": 0.1224,
+      "step": 18190
+    },
+    {
+      "epoch": 16.33,
+      "learning_rate": 3.7402752842609215e-05,
+      "loss": 0.1136,
+      "step": 18200
+    },
+    {
+      "epoch": 16.34,
+      "learning_rate": 3.739277877518452e-05,
+      "loss": 0.1146,
+      "step": 18210
+    },
+    {
+      "epoch": 16.34,
+      "learning_rate": 3.738280470775983e-05,
+      "loss": 0.1118,
+      "step": 18220
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 3.7372830640335134e-05,
+      "loss": 0.1076,
+      "step": 18230
+    },
+    {
+      "epoch": 16.36,
+      "learning_rate": 3.736285657291044e-05,
+      "loss": 0.0857,
+      "step": 18240
+    },
+    {
+      "epoch": 16.37,
+      "learning_rate": 3.735288250548574e-05,
+      "loss": 0.0925,
+      "step": 18250
+    },
+    {
+      "epoch": 16.38,
+      "learning_rate": 3.734290843806104e-05,
+      "loss": 0.111,
+      "step": 18260
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 3.7332934370636345e-05,
+      "loss": 0.0695,
+      "step": 18270
+    },
+    {
+      "epoch": 16.4,
+      "learning_rate": 3.732296030321165e-05,
+      "loss": 0.0774,
+      "step": 18280
+    },
+    {
+      "epoch": 16.41,
+      "learning_rate": 3.731298623578696e-05,
+      "loss": 0.1057,
+      "step": 18290
+    },
+    {
+      "epoch": 16.42,
+      "learning_rate": 3.7303012168362264e-05,
+      "loss": 0.0812,
+      "step": 18300
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 3.729303810093756e-05,
+      "loss": 0.0735,
+      "step": 18310
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 3.728306403351286e-05,
+      "loss": 0.0934,
+      "step": 18320
+    },
+    {
+      "epoch": 16.44,
+      "learning_rate": 3.727308996608817e-05,
+      "loss": 0.0882,
+      "step": 18330
+    },
+    {
+      "epoch": 16.45,
+      "learning_rate": 3.7263115898663475e-05,
+      "loss": 0.1118,
+      "step": 18340
+    },
+    {
+      "epoch": 16.46,
+      "learning_rate": 3.725314183123878e-05,
+      "loss": 0.0736,
+      "step": 18350
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 3.724316776381409e-05,
+      "loss": 0.0879,
+      "step": 18360
+    },
+    {
+      "epoch": 16.48,
+      "learning_rate": 3.723319369638939e-05,
+      "loss": 0.0929,
+      "step": 18370
+    },
+    {
+      "epoch": 16.49,
+      "learning_rate": 3.722321962896469e-05,
+      "loss": 0.0748,
+      "step": 18380
+    },
+    {
+      "epoch": 16.5,
+      "learning_rate": 3.721324556154e-05,
+      "loss": 0.0813,
+      "step": 18390
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 3.72032714941153e-05,
+      "loss": 0.0866,
+      "step": 18400
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 3.7193297426690605e-05,
+      "loss": 0.0897,
+      "step": 18410
+    },
+    {
+      "epoch": 16.52,
+      "learning_rate": 3.718332335926591e-05,
+      "loss": 0.1175,
+      "step": 18420
+    },
+    {
+      "epoch": 16.53,
+      "learning_rate": 3.717334929184122e-05,
+      "loss": 0.1085,
+      "step": 18430
+    },
+    {
+      "epoch": 16.54,
+      "learning_rate": 3.7163375224416516e-05,
+      "loss": 0.1173,
+      "step": 18440
+    },
+    {
+      "epoch": 16.55,
+      "learning_rate": 3.715340115699182e-05,
+      "loss": 0.0678,
+      "step": 18450
+    },
+    {
+      "epoch": 16.56,
+      "learning_rate": 3.714342708956713e-05,
+      "loss": 0.0691,
+      "step": 18460
+    },
+    {
+      "epoch": 16.57,
+      "learning_rate": 3.7133453022142435e-05,
+      "loss": 0.1089,
+      "step": 18470
+    },
+    {
+      "epoch": 16.58,
+      "learning_rate": 3.7123478954717734e-05,
+      "loss": 0.0936,
+      "step": 18480
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 3.711350488729304e-05,
+      "loss": 0.1011,
+      "step": 18490
+    },
+    {
+      "epoch": 16.6,
+      "learning_rate": 3.710353081986834e-05,
+      "loss": 0.0923,
+      "step": 18500
+    },
+    {
+      "epoch": 16.6,
+      "learning_rate": 3.7093556752443646e-05,
+      "loss": 0.1171,
+      "step": 18510
+    },
+    {
+      "epoch": 16.61,
+      "learning_rate": 3.708358268501895e-05,
+      "loss": 0.0964,
+      "step": 18520
+    },
+    {
+      "epoch": 16.62,
+      "learning_rate": 3.707360861759426e-05,
+      "loss": 0.0862,
+      "step": 18530
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 3.7063634550169565e-05,
+      "loss": 0.1237,
+      "step": 18540
+    },
+    {
+      "epoch": 16.64,
+      "learning_rate": 3.7053660482744864e-05,
+      "loss": 0.0903,
+      "step": 18550
+    },
+    {
+      "epoch": 16.65,
+      "learning_rate": 3.7043686415320164e-05,
+      "loss": 0.1073,
+      "step": 18560
+    },
+    {
+      "epoch": 16.66,
+      "learning_rate": 3.703371234789547e-05,
+      "loss": 0.0738,
+      "step": 18570
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 3.7023738280470776e-05,
+      "loss": 0.1056,
+      "step": 18580
+    },
+    {
+      "epoch": 16.68,
+      "learning_rate": 3.701376421304608e-05,
+      "loss": 0.0998,
+      "step": 18590
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 3.700379014562139e-05,
+      "loss": 0.11,
+      "step": 18600
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 3.6993816078196694e-05,
+      "loss": 0.1278,
+      "step": 18610
+    },
+    {
+      "epoch": 16.7,
+      "learning_rate": 3.6983842010771994e-05,
+      "loss": 0.1024,
+      "step": 18620
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 3.69738679433473e-05,
+      "loss": 0.0947,
+      "step": 18630
+    },
+    {
+      "epoch": 16.72,
+      "learning_rate": 3.69638938759226e-05,
+      "loss": 0.0887,
+      "step": 18640
+    },
+    {
+      "epoch": 16.73,
+      "learning_rate": 3.6953919808497906e-05,
+      "loss": 0.0854,
+      "step": 18650
+    },
+    {
+      "epoch": 16.74,
+      "learning_rate": 3.694394574107321e-05,
+      "loss": 0.0797,
+      "step": 18660
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 3.693397167364852e-05,
+      "loss": 0.0624,
+      "step": 18670
+    },
+    {
+      "epoch": 16.76,
+      "learning_rate": 3.692399760622382e-05,
+      "loss": 0.0935,
+      "step": 18680
+    },
+    {
+      "epoch": 16.77,
+      "learning_rate": 3.6914023538799124e-05,
+      "loss": 0.0711,
+      "step": 18690
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 3.690404947137443e-05,
+      "loss": 0.0784,
+      "step": 18700
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 3.6894075403949736e-05,
+      "loss": 0.112,
+      "step": 18710
+    },
+    {
+      "epoch": 16.79,
+      "learning_rate": 3.6884101336525035e-05,
+      "loss": 0.0624,
+      "step": 18720
+    },
+    {
+      "epoch": 16.8,
+      "learning_rate": 3.687412726910034e-05,
+      "loss": 0.0772,
+      "step": 18730
+    },
+    {
+      "epoch": 16.81,
+      "learning_rate": 3.686415320167564e-05,
+      "loss": 0.1138,
+      "step": 18740
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 3.685417913425095e-05,
+      "loss": 0.1076,
+      "step": 18750
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 3.6844205066826253e-05,
+      "loss": 0.0565,
+      "step": 18760
+    },
+    {
+      "epoch": 16.84,
+      "learning_rate": 3.683423099940156e-05,
+      "loss": 0.0906,
+      "step": 18770
+    },
+    {
+      "epoch": 16.85,
+      "learning_rate": 3.6824256931976866e-05,
+      "loss": 0.0892,
+      "step": 18780
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 3.681428286455217e-05,
+      "loss": 0.1184,
+      "step": 18790
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 3.6804308797127465e-05,
+      "loss": 0.0929,
+      "step": 18800
+    },
+    {
+      "epoch": 16.87,
+      "learning_rate": 3.679433472970277e-05,
+      "loss": 0.0755,
+      "step": 18810
+    },
+    {
+      "epoch": 16.88,
+      "learning_rate": 3.678436066227808e-05,
+      "loss": 0.098,
+      "step": 18820
+    },
+    {
+      "epoch": 16.89,
+      "learning_rate": 3.677438659485338e-05,
+      "loss": 0.0588,
+      "step": 18830
+    },
+    {
+      "epoch": 16.9,
+      "learning_rate": 3.676441252742869e-05,
+      "loss": 0.1004,
+      "step": 18840
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 3.6754438460003996e-05,
+      "loss": 0.0914,
+      "step": 18850
+    },
+    {
+      "epoch": 16.92,
+      "learning_rate": 3.6744464392579295e-05,
+      "loss": 0.0611,
+      "step": 18860
+    },
+    {
+      "epoch": 16.93,
+      "learning_rate": 3.67344903251546e-05,
+      "loss": 0.1051,
+      "step": 18870
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 3.67245162577299e-05,
+      "loss": 0.108,
+      "step": 18880
+    },
+    {
+      "epoch": 16.95,
+      "learning_rate": 3.671454219030521e-05,
+      "loss": 0.0719,
+      "step": 18890
+    },
+    {
+      "epoch": 16.95,
+      "learning_rate": 3.670456812288051e-05,
+      "loss": 0.0678,
+      "step": 18900
+    },
+    {
+      "epoch": 16.96,
+      "learning_rate": 3.669459405545582e-05,
+      "loss": 0.1023,
+      "step": 18910
+    },
+    {
+      "epoch": 16.97,
+      "learning_rate": 3.668461998803112e-05,
+      "loss": 0.1111,
+      "step": 18920
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 3.6674645920606425e-05,
+      "loss": 0.12,
+      "step": 18930
+    },
+    {
+      "epoch": 16.99,
+      "learning_rate": 3.666467185318173e-05,
+      "loss": 0.0928,
+      "step": 18940
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 3.665469778575704e-05,
+      "loss": 0.1055,
+      "step": 18950
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": {
+        "accuracy": 0.9753307916573223
+      },
+      "eval_f1": {
+        "f1": 0.9739943140941046
+      },
+      "eval_loss": 0.07092935591936111,
+      "eval_precision": {
+        "precision": 0.974174353159638
+      },
+      "eval_recall": {
+        "recall": 0.9738453333404872
+      },
+      "eval_runtime": 198.296,
+      "eval_samples_per_second": 89.946,
+      "eval_steps_per_second": 5.623,
+      "step": 18950
+    },
+    {
+      "epoch": 17.01,
+      "learning_rate": 3.664472371833234e-05,
+      "loss": 0.0713,
+      "step": 18960
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 3.663474965090764e-05,
+      "loss": 0.0594,
+      "step": 18970
+    },
+    {
+      "epoch": 17.03,
+      "learning_rate": 3.662477558348294e-05,
+      "loss": 0.0647,
+      "step": 18980
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 3.661480151605825e-05,
+      "loss": 0.0899,
+      "step": 18990
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 3.6604827448633555e-05,
+      "loss": 0.0932,
+      "step": 19000
+    },
+    {
+      "epoch": 17.05,
+      "learning_rate": 3.659485338120886e-05,
+      "loss": 0.086,
+      "step": 19010
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 3.658487931378417e-05,
+      "loss": 0.0783,
+      "step": 19020
+    },
+    {
+      "epoch": 17.07,
+      "learning_rate": 3.657490524635947e-05,
+      "loss": 0.0787,
+      "step": 19030
+    },
+    {
+      "epoch": 17.08,
+      "learning_rate": 3.656493117893477e-05,
+      "loss": 0.0719,
+      "step": 19040
+    },
+    {
+      "epoch": 17.09,
+      "learning_rate": 3.655495711151007e-05,
+      "loss": 0.1086,
+      "step": 19050
+    },
+    {
+      "epoch": 17.1,
+      "learning_rate": 3.654498304408538e-05,
+      "loss": 0.0732,
+      "step": 19060
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 3.6535008976660684e-05,
+      "loss": 0.0959,
+      "step": 19070
+    },
+    {
+      "epoch": 17.12,
+      "learning_rate": 3.652503490923599e-05,
+      "loss": 0.064,
+      "step": 19080
+    },
+    {
+      "epoch": 17.12,
+      "learning_rate": 3.65150608418113e-05,
+      "loss": 0.1165,
+      "step": 19090
+    },
+    {
+      "epoch": 17.13,
+      "learning_rate": 3.6505086774386596e-05,
+      "loss": 0.0916,
+      "step": 19100
+    },
+    {
+      "epoch": 17.14,
+      "learning_rate": 3.6495112706961896e-05,
+      "loss": 0.0973,
+      "step": 19110
+    },
+    {
+      "epoch": 17.15,
+      "learning_rate": 3.64851386395372e-05,
+      "loss": 0.0991,
+      "step": 19120
+    },
+    {
+      "epoch": 17.16,
+      "learning_rate": 3.647516457211251e-05,
+      "loss": 0.0868,
+      "step": 19130
+    },
+    {
+      "epoch": 17.17,
+      "learning_rate": 3.6465190504687814e-05,
+      "loss": 0.0832,
+      "step": 19140
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 3.645521643726312e-05,
+      "loss": 0.0947,
+      "step": 19150
+    },
+    {
+      "epoch": 17.19,
+      "learning_rate": 3.644524236983842e-05,
+      "loss": 0.0963,
+      "step": 19160
+    },
+    {
+      "epoch": 17.2,
+      "learning_rate": 3.6435268302413726e-05,
+      "loss": 0.0816,
+      "step": 19170
+    },
+    {
+      "epoch": 17.21,
+      "learning_rate": 3.642529423498903e-05,
+      "loss": 0.0879,
+      "step": 19180
+    },
+    {
+      "epoch": 17.21,
+      "learning_rate": 3.641532016756433e-05,
+      "loss": 0.0775,
+      "step": 19190
+    },
+    {
+      "epoch": 17.22,
+      "learning_rate": 3.640534610013964e-05,
+      "loss": 0.0619,
+      "step": 19200
+    },
+    {
+      "epoch": 17.23,
+      "learning_rate": 3.6395372032714944e-05,
+      "loss": 0.0991,
+      "step": 19210
+    },
+    {
+      "epoch": 17.24,
+      "learning_rate": 3.6385397965290243e-05,
+      "loss": 0.0999,
+      "step": 19220
+    },
+    {
+      "epoch": 17.25,
+      "learning_rate": 3.637542389786555e-05,
+      "loss": 0.1133,
+      "step": 19230
+    },
+    {
+      "epoch": 17.26,
+      "learning_rate": 3.6365449830440856e-05,
+      "loss": 0.1232,
+      "step": 19240
+    },
+    {
+      "epoch": 17.27,
+      "learning_rate": 3.635547576301616e-05,
+      "loss": 0.0939,
+      "step": 19250
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 3.634550169559147e-05,
+      "loss": 0.087,
+      "step": 19260
+    },
+    {
+      "epoch": 17.29,
+      "learning_rate": 3.633552762816677e-05,
+      "loss": 0.0965,
+      "step": 19270
+    },
+    {
+      "epoch": 17.3,
+      "learning_rate": 3.6325553560742074e-05,
+      "loss": 0.0853,
+      "step": 19280
+    },
+    {
+      "epoch": 17.3,
+      "learning_rate": 3.631557949331737e-05,
+      "loss": 0.0922,
+      "step": 19290
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 3.630560542589268e-05,
+      "loss": 0.0943,
+      "step": 19300
+    },
+    {
+      "epoch": 17.32,
+      "learning_rate": 3.6295631358467986e-05,
+      "loss": 0.0862,
+      "step": 19310
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 3.628565729104329e-05,
+      "loss": 0.0954,
+      "step": 19320
+    },
+    {
+      "epoch": 17.34,
+      "learning_rate": 3.62756832236186e-05,
+      "loss": 0.0922,
+      "step": 19330
+    },
+    {
+      "epoch": 17.35,
+      "learning_rate": 3.62657091561939e-05,
+      "loss": 0.0616,
+      "step": 19340
+    },
+    {
+      "epoch": 17.36,
+      "learning_rate": 3.62557350887692e-05,
+      "loss": 0.0813,
+      "step": 19350
+    },
+    {
+      "epoch": 17.37,
+      "learning_rate": 3.62457610213445e-05,
+      "loss": 0.0784,
+      "step": 19360
+    },
+    {
+      "epoch": 17.38,
+      "learning_rate": 3.623578695391981e-05,
+      "loss": 0.0776,
+      "step": 19370
+    },
+    {
+      "epoch": 17.39,
+      "learning_rate": 3.6225812886495115e-05,
+      "loss": 0.088,
+      "step": 19380
+    },
+    {
+      "epoch": 17.39,
+      "learning_rate": 3.621583881907042e-05,
+      "loss": 0.0857,
+      "step": 19390
+    },
+    {
+      "epoch": 17.4,
+      "learning_rate": 3.620586475164572e-05,
+      "loss": 0.0757,
+      "step": 19400
+    },
+    {
+      "epoch": 17.41,
+      "learning_rate": 3.619589068422103e-05,
+      "loss": 0.0998,
+      "step": 19410
+    },
+    {
+      "epoch": 17.42,
+      "learning_rate": 3.618591661679633e-05,
+      "loss": 0.0798,
+      "step": 19420
+    },
+    {
+      "epoch": 17.43,
+      "learning_rate": 3.617594254937163e-05,
+      "loss": 0.11,
+      "step": 19430
+    },
+    {
+      "epoch": 17.44,
+      "learning_rate": 3.616596848194694e-05,
+      "loss": 0.0862,
+      "step": 19440
+    },
+    {
+      "epoch": 17.45,
+      "learning_rate": 3.6155994414522245e-05,
+      "loss": 0.068,
+      "step": 19450
+    },
+    {
+      "epoch": 17.46,
+      "learning_rate": 3.6146020347097545e-05,
+      "loss": 0.0866,
+      "step": 19460
+    },
+    {
+      "epoch": 17.47,
+      "learning_rate": 3.613604627967285e-05,
+      "loss": 0.1017,
+      "step": 19470
+    },
+    {
+      "epoch": 17.47,
+      "learning_rate": 3.612607221224816e-05,
+      "loss": 0.0806,
+      "step": 19480
+    },
+    {
+      "epoch": 17.48,
+      "learning_rate": 3.611609814482346e-05,
+      "loss": 0.1168,
+      "step": 19490
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 3.610612407739877e-05,
+      "loss": 0.0593,
+      "step": 19500
+    },
+    {
+      "epoch": 17.5,
+      "learning_rate": 3.609615000997407e-05,
+      "loss": 0.1227,
+      "step": 19510
+    },
+    {
+      "epoch": 17.51,
+      "learning_rate": 3.6086175942549375e-05,
+      "loss": 0.0803,
+      "step": 19520
+    },
+    {
+      "epoch": 17.52,
+      "learning_rate": 3.6076201875124674e-05,
+      "loss": 0.0972,
+      "step": 19530
+    },
+    {
+      "epoch": 17.53,
+      "learning_rate": 3.606622780769998e-05,
+      "loss": 0.0716,
+      "step": 19540
+    },
+    {
+      "epoch": 17.54,
+      "learning_rate": 3.605625374027529e-05,
+      "loss": 0.0897,
+      "step": 19550
+    },
+    {
+      "epoch": 17.55,
+      "learning_rate": 3.604627967285059e-05,
+      "loss": 0.1333,
+      "step": 19560
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 3.60363056054259e-05,
+      "loss": 0.0495,
+      "step": 19570
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 3.60263315380012e-05,
+      "loss": 0.093,
+      "step": 19580
+    },
+    {
+      "epoch": 17.57,
+      "learning_rate": 3.60163574705765e-05,
+      "loss": 0.0806,
+      "step": 19590
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 3.6006383403151804e-05,
+      "loss": 0.1035,
+      "step": 19600
+    },
+    {
+      "epoch": 17.59,
+      "learning_rate": 3.599640933572711e-05,
+      "loss": 0.076,
+      "step": 19610
+    },
+    {
+      "epoch": 17.6,
+      "learning_rate": 3.5986435268302417e-05,
+      "loss": 0.1181,
+      "step": 19620
+    },
+    {
+      "epoch": 17.61,
+      "learning_rate": 3.597646120087772e-05,
+      "loss": 0.0861,
+      "step": 19630
+    },
+    {
+      "epoch": 17.62,
+      "learning_rate": 3.596648713345302e-05,
+      "loss": 0.0722,
+      "step": 19640
+    },
+    {
+      "epoch": 17.63,
+      "learning_rate": 3.595651306602833e-05,
+      "loss": 0.0907,
+      "step": 19650
+    },
+    {
+      "epoch": 17.64,
+      "learning_rate": 3.5946538998603635e-05,
+      "loss": 0.0951,
+      "step": 19660
+    },
+    {
+      "epoch": 17.65,
+      "learning_rate": 3.5936564931178934e-05,
+      "loss": 0.0866,
+      "step": 19670
+    },
+    {
+      "epoch": 17.65,
+      "learning_rate": 3.592659086375424e-05,
+      "loss": 0.0913,
+      "step": 19680
+    },
+    {
+      "epoch": 17.66,
+      "learning_rate": 3.5916616796329546e-05,
+      "loss": 0.108,
+      "step": 19690
+    },
+    {
+      "epoch": 17.67,
+      "learning_rate": 3.5906642728904846e-05,
+      "loss": 0.1095,
+      "step": 19700
+    },
+    {
+      "epoch": 17.68,
+      "learning_rate": 3.589666866148015e-05,
+      "loss": 0.0733,
+      "step": 19710
+    },
+    {
+      "epoch": 17.69,
+      "learning_rate": 3.588669459405546e-05,
+      "loss": 0.0739,
+      "step": 19720
+    },
+    {
+      "epoch": 17.7,
+      "learning_rate": 3.5876720526630764e-05,
+      "loss": 0.111,
+      "step": 19730
+    },
+    {
+      "epoch": 17.71,
+      "learning_rate": 3.586674645920607e-05,
+      "loss": 0.0876,
+      "step": 19740
+    },
+    {
+      "epoch": 17.72,
+      "learning_rate": 3.585677239178137e-05,
+      "loss": 0.1067,
+      "step": 19750
+    },
+    {
+      "epoch": 17.73,
+      "learning_rate": 3.5846798324356676e-05,
+      "loss": 0.1034,
+      "step": 19760
+    },
+    {
+      "epoch": 17.73,
+      "learning_rate": 3.5836824256931976e-05,
+      "loss": 0.1041,
+      "step": 19770
+    },
+    {
+      "epoch": 17.74,
+      "learning_rate": 3.582685018950728e-05,
+      "loss": 0.0712,
+      "step": 19780
+    },
+    {
+      "epoch": 17.75,
+      "learning_rate": 3.581687612208259e-05,
+      "loss": 0.0574,
+      "step": 19790
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 3.5806902054657894e-05,
+      "loss": 0.1125,
+      "step": 19800
+    },
+    {
+      "epoch": 17.77,
+      "learning_rate": 3.57969279872332e-05,
+      "loss": 0.1031,
+      "step": 19810
+    },
+    {
+      "epoch": 17.78,
+      "learning_rate": 3.57869539198085e-05,
+      "loss": 0.0723,
+      "step": 19820
+    },
+    {
+      "epoch": 17.79,
+      "learning_rate": 3.57769798523838e-05,
+      "loss": 0.1072,
+      "step": 19830
+    },
+    {
+      "epoch": 17.8,
+      "learning_rate": 3.5767005784959105e-05,
+      "loss": 0.1189,
+      "step": 19840
+    },
+    {
+      "epoch": 17.81,
+      "learning_rate": 3.575703171753441e-05,
+      "loss": 0.0938,
+      "step": 19850
+    },
+    {
+      "epoch": 17.82,
+      "learning_rate": 3.574705765010972e-05,
+      "loss": 0.0888,
+      "step": 19860
+    },
+    {
+      "epoch": 17.82,
+      "learning_rate": 3.5737083582685024e-05,
+      "loss": 0.0527,
+      "step": 19870
+    },
+    {
+      "epoch": 17.83,
+      "learning_rate": 3.572710951526032e-05,
+      "loss": 0.1048,
+      "step": 19880
+    },
+    {
+      "epoch": 17.84,
+      "learning_rate": 3.571713544783563e-05,
+      "loss": 0.0998,
+      "step": 19890
+    },
+    {
+      "epoch": 17.85,
+      "learning_rate": 3.570716138041093e-05,
+      "loss": 0.0877,
+      "step": 19900
+    },
+    {
+      "epoch": 17.86,
+      "learning_rate": 3.5697187312986235e-05,
+      "loss": 0.0723,
+      "step": 19910
+    },
+    {
+      "epoch": 17.87,
+      "learning_rate": 3.568721324556154e-05,
+      "loss": 0.1051,
+      "step": 19920
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 3.567723917813685e-05,
+      "loss": 0.0899,
+      "step": 19930
+    },
+    {
+      "epoch": 17.89,
+      "learning_rate": 3.566726511071215e-05,
+      "loss": 0.0736,
+      "step": 19940
+    },
+    {
+      "epoch": 17.9,
+      "learning_rate": 3.565729104328745e-05,
+      "loss": 0.071,
+      "step": 19950
+    },
+    {
+      "epoch": 17.91,
+      "learning_rate": 3.564731697586276e-05,
+      "loss": 0.0671,
+      "step": 19960
+    },
+    {
+      "epoch": 17.91,
+      "learning_rate": 3.5637342908438065e-05,
+      "loss": 0.0979,
+      "step": 19970
+    },
+    {
+      "epoch": 17.92,
+      "learning_rate": 3.5627368841013365e-05,
+      "loss": 0.1133,
+      "step": 19980
+    },
+    {
+      "epoch": 17.93,
+      "learning_rate": 3.561739477358867e-05,
+      "loss": 0.0683,
+      "step": 19990
+    },
+    {
+      "epoch": 17.94,
+      "learning_rate": 3.560742070616398e-05,
+      "loss": 0.0993,
+      "step": 20000
+    },
+    {
+      "epoch": 17.95,
+      "learning_rate": 3.559744663873928e-05,
+      "loss": 0.0966,
+      "step": 20010
+    },
+    {
+      "epoch": 17.96,
+      "learning_rate": 3.558747257131458e-05,
+      "loss": 0.0637,
+      "step": 20020
+    },
+    {
+      "epoch": 17.97,
+      "learning_rate": 3.557749850388989e-05,
+      "loss": 0.0854,
+      "step": 20030
+    },
+    {
+      "epoch": 17.98,
+      "learning_rate": 3.5567524436465195e-05,
+      "loss": 0.0679,
+      "step": 20040
+    },
+    {
+      "epoch": 17.99,
+      "learning_rate": 3.55575503690405e-05,
+      "loss": 0.1121,
+      "step": 20050
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 3.55475763016158e-05,
+      "loss": 0.0644,
+      "step": 20060
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": {
+        "accuracy": 0.9766763848396501
+      },
+      "eval_f1": {
+        "f1": 0.9755727076296268
+      },
+      "eval_loss": 0.07384093105792999,
+      "eval_precision": {
+        "precision": 0.9750313848950637
+      },
+      "eval_recall": {
+        "recall": 0.9764061086777133
+      },
+      "eval_runtime": 185.6066,
+      "eval_samples_per_second": 96.096,
+      "eval_steps_per_second": 6.007,
+      "step": 20065
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 3.55376022341911e-05,
+      "loss": 0.0931,
+      "step": 20070
+    },
+    {
+      "epoch": 18.01,
+      "learning_rate": 3.5527628166766406e-05,
+      "loss": 0.0785,
+      "step": 20080
+    },
+    {
+      "epoch": 18.02,
+      "learning_rate": 3.551765409934171e-05,
+      "loss": 0.093,
+      "step": 20090
+    },
+    {
+      "epoch": 18.03,
+      "learning_rate": 3.550768003191702e-05,
+      "loss": 0.0793,
+      "step": 20100
+    },
+    {
+      "epoch": 18.04,
+      "learning_rate": 3.5497705964492325e-05,
+      "loss": 0.0964,
+      "step": 20110
+    },
+    {
+      "epoch": 18.05,
+      "learning_rate": 3.5487731897067624e-05,
+      "loss": 0.0809,
+      "step": 20120
+    },
+    {
+      "epoch": 18.06,
+      "learning_rate": 3.547775782964293e-05,
+      "loss": 0.0839,
+      "step": 20130
+    },
+    {
+      "epoch": 18.07,
+      "learning_rate": 3.546778376221823e-05,
+      "loss": 0.0716,
+      "step": 20140
+    },
+    {
+      "epoch": 18.08,
+      "learning_rate": 3.5457809694793536e-05,
+      "loss": 0.0569,
+      "step": 20150
+    },
+    {
+      "epoch": 18.08,
+      "learning_rate": 3.544783562736884e-05,
+      "loss": 0.1029,
+      "step": 20160
+    },
+    {
+      "epoch": 18.09,
+      "learning_rate": 3.543786155994415e-05,
+      "loss": 0.0989,
+      "step": 20170
+    },
+    {
+      "epoch": 18.1,
+      "learning_rate": 3.5427887492519455e-05,
+      "loss": 0.083,
+      "step": 20180
+    },
+    {
+      "epoch": 18.11,
+      "learning_rate": 3.5417913425094754e-05,
+      "loss": 0.0778,
+      "step": 20190
+    },
+    {
+      "epoch": 18.12,
+      "learning_rate": 3.540793935767006e-05,
+      "loss": 0.0899,
+      "step": 20200
+    },
+    {
+      "epoch": 18.13,
+      "learning_rate": 3.539796529024537e-05,
+      "loss": 0.0676,
+      "step": 20210
+    },
+    {
+      "epoch": 18.14,
+      "learning_rate": 3.5387991222820666e-05,
+      "loss": 0.0978,
+      "step": 20220
+    },
+    {
+      "epoch": 18.15,
+      "learning_rate": 3.537801715539597e-05,
+      "loss": 0.0878,
+      "step": 20230
+    },
+    {
+      "epoch": 18.16,
+      "learning_rate": 3.536804308797128e-05,
+      "loss": 0.0849,
+      "step": 20240
+    },
+    {
+      "epoch": 18.17,
+      "learning_rate": 3.535806902054658e-05,
+      "loss": 0.0696,
+      "step": 20250
+    },
+    {
+      "epoch": 18.17,
+      "learning_rate": 3.5348094953121884e-05,
+      "loss": 0.0564,
+      "step": 20260
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 3.533812088569719e-05,
+      "loss": 0.0825,
+      "step": 20270
+    },
+    {
+      "epoch": 18.19,
+      "learning_rate": 3.5328146818272496e-05,
+      "loss": 0.0844,
+      "step": 20280
+    },
+    {
+      "epoch": 18.2,
+      "learning_rate": 3.53181727508478e-05,
+      "loss": 0.0784,
+      "step": 20290
+    },
+    {
+      "epoch": 18.21,
+      "learning_rate": 3.53081986834231e-05,
+      "loss": 0.0912,
+      "step": 20300
+    },
+    {
+      "epoch": 18.22,
+      "learning_rate": 3.52982246159984e-05,
+      "loss": 0.1091,
+      "step": 20310
+    },
+    {
+      "epoch": 18.23,
+      "learning_rate": 3.528825054857371e-05,
+      "loss": 0.0692,
+      "step": 20320
+    },
+    {
+      "epoch": 18.24,
+      "learning_rate": 3.5278276481149014e-05,
+      "loss": 0.0738,
+      "step": 20330
+    },
+    {
+      "epoch": 18.25,
+      "learning_rate": 3.526830241372432e-05,
+      "loss": 0.1141,
+      "step": 20340
+    },
+    {
+      "epoch": 18.26,
+      "learning_rate": 3.5258328346299626e-05,
+      "loss": 0.0571,
+      "step": 20350
+    },
+    {
+      "epoch": 18.26,
+      "learning_rate": 3.5248354278874926e-05,
+      "loss": 0.0952,
+      "step": 20360
+    },
+    {
+      "epoch": 18.27,
+      "learning_rate": 3.523838021145023e-05,
+      "loss": 0.0732,
+      "step": 20370
+    },
+    {
+      "epoch": 18.28,
+      "learning_rate": 3.522840614402553e-05,
+      "loss": 0.0801,
+      "step": 20380
+    },
+    {
+      "epoch": 18.29,
+      "learning_rate": 3.521843207660084e-05,
+      "loss": 0.1099,
+      "step": 20390
+    },
+    {
+      "epoch": 18.3,
+      "learning_rate": 3.5208458009176144e-05,
+      "loss": 0.0743,
+      "step": 20400
+    },
+    {
+      "epoch": 18.31,
+      "learning_rate": 3.519848394175145e-05,
+      "loss": 0.0755,
+      "step": 20410
+    },
+    {
+      "epoch": 18.32,
+      "learning_rate": 3.5188509874326756e-05,
+      "loss": 0.087,
+      "step": 20420
+    },
+    {
+      "epoch": 18.33,
+      "learning_rate": 3.5178535806902055e-05,
+      "loss": 0.0929,
+      "step": 20430
+    },
+    {
+      "epoch": 18.34,
+      "learning_rate": 3.516856173947736e-05,
+      "loss": 0.0744,
+      "step": 20440
+    },
+    {
+      "epoch": 18.34,
+      "learning_rate": 3.515858767205267e-05,
+      "loss": 0.0647,
+      "step": 20450
+    },
+    {
+      "epoch": 18.35,
+      "learning_rate": 3.514861360462797e-05,
+      "loss": 0.065,
+      "step": 20460
+    },
+    {
+      "epoch": 18.36,
+      "learning_rate": 3.5138639537203273e-05,
+      "loss": 0.1091,
+      "step": 20470
+    },
+    {
+      "epoch": 18.37,
+      "learning_rate": 3.512866546977858e-05,
+      "loss": 0.0718,
+      "step": 20480
+    },
+    {
+      "epoch": 18.38,
+      "learning_rate": 3.511869140235388e-05,
+      "loss": 0.0729,
+      "step": 20490
+    },
+    {
+      "epoch": 18.39,
+      "learning_rate": 3.5108717334929185e-05,
+      "loss": 0.0882,
+      "step": 20500
+    },
+    {
+      "epoch": 18.4,
+      "learning_rate": 3.509874326750449e-05,
+      "loss": 0.0724,
+      "step": 20510
+    },
+    {
+      "epoch": 18.41,
+      "learning_rate": 3.50887692000798e-05,
+      "loss": 0.0942,
+      "step": 20520
+    },
+    {
+      "epoch": 18.42,
+      "learning_rate": 3.5078795132655104e-05,
+      "loss": 0.0722,
+      "step": 20530
+    },
+    {
+      "epoch": 18.43,
+      "learning_rate": 3.50688210652304e-05,
+      "loss": 0.0762,
+      "step": 20540
+    },
+    {
+      "epoch": 18.43,
+      "learning_rate": 3.50588469978057e-05,
+      "loss": 0.0917,
+      "step": 20550
+    },
+    {
+      "epoch": 18.44,
+      "learning_rate": 3.504887293038101e-05,
+      "loss": 0.0665,
+      "step": 20560
+    },
+    {
+      "epoch": 18.45,
+      "learning_rate": 3.5038898862956315e-05,
+      "loss": 0.091,
+      "step": 20570
+    },
+    {
+      "epoch": 18.46,
+      "learning_rate": 3.502892479553162e-05,
+      "loss": 0.079,
+      "step": 20580
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 3.501895072810693e-05,
+      "loss": 0.0608,
+      "step": 20590
+    },
+    {
+      "epoch": 18.48,
+      "learning_rate": 3.500897666068223e-05,
+      "loss": 0.0843,
+      "step": 20600
+    },
+    {
+      "epoch": 18.49,
+      "learning_rate": 3.4999002593257526e-05,
+      "loss": 0.0741,
+      "step": 20610
+    },
+    {
+      "epoch": 18.5,
+      "learning_rate": 3.498902852583283e-05,
+      "loss": 0.0933,
+      "step": 20620
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 3.497905445840814e-05,
+      "loss": 0.0435,
+      "step": 20630
+    },
+    {
+      "epoch": 18.52,
+      "learning_rate": 3.4969080390983445e-05,
+      "loss": 0.0906,
+      "step": 20640
+    },
+    {
+      "epoch": 18.52,
+      "learning_rate": 3.495910632355875e-05,
+      "loss": 0.109,
+      "step": 20650
+    },
+    {
+      "epoch": 18.53,
+      "learning_rate": 3.494913225613406e-05,
+      "loss": 0.0641,
+      "step": 20660
+    },
+    {
+      "epoch": 18.54,
+      "learning_rate": 3.4939158188709357e-05,
+      "loss": 0.0738,
+      "step": 20670
+    },
+    {
+      "epoch": 18.55,
+      "learning_rate": 3.492918412128466e-05,
+      "loss": 0.0664,
+      "step": 20680
+    },
+    {
+      "epoch": 18.56,
+      "learning_rate": 3.491921005385996e-05,
+      "loss": 0.1015,
+      "step": 20690
+    },
+    {
+      "epoch": 18.57,
+      "learning_rate": 3.490923598643527e-05,
+      "loss": 0.0647,
+      "step": 20700
+    },
+    {
+      "epoch": 18.58,
+      "learning_rate": 3.4899261919010575e-05,
+      "loss": 0.1174,
+      "step": 20710
+    },
+    {
+      "epoch": 18.59,
+      "learning_rate": 3.488928785158588e-05,
+      "loss": 0.0953,
+      "step": 20720
+    },
+    {
+      "epoch": 18.6,
+      "learning_rate": 3.487931378416118e-05,
+      "loss": 0.08,
+      "step": 20730
+    },
+    {
+      "epoch": 18.61,
+      "learning_rate": 3.4869339716736486e-05,
+      "loss": 0.091,
+      "step": 20740
+    },
+    {
+      "epoch": 18.61,
+      "learning_rate": 3.485936564931179e-05,
+      "loss": 0.0719,
+      "step": 20750
+    },
+    {
+      "epoch": 18.62,
+      "learning_rate": 3.48493915818871e-05,
+      "loss": 0.0702,
+      "step": 20760
+    },
+    {
+      "epoch": 18.63,
+      "learning_rate": 3.48394175144624e-05,
+      "loss": 0.0892,
+      "step": 20770
+    },
+    {
+      "epoch": 18.64,
+      "learning_rate": 3.4829443447037704e-05,
+      "loss": 0.0523,
+      "step": 20780
+    },
+    {
+      "epoch": 18.65,
+      "learning_rate": 3.4819469379613004e-05,
+      "loss": 0.0458,
+      "step": 20790
+    },
+    {
+      "epoch": 18.66,
+      "learning_rate": 3.480949531218831e-05,
+      "loss": 0.1033,
+      "step": 20800
+    },
+    {
+      "epoch": 18.67,
+      "learning_rate": 3.4799521244763616e-05,
+      "loss": 0.0801,
+      "step": 20810
+    },
+    {
+      "epoch": 18.68,
+      "learning_rate": 3.478954717733892e-05,
+      "loss": 0.1108,
+      "step": 20820
+    },
+    {
+      "epoch": 18.69,
+      "learning_rate": 3.477957310991423e-05,
+      "loss": 0.1124,
+      "step": 20830
+    },
+    {
+      "epoch": 18.69,
+      "learning_rate": 3.476959904248953e-05,
+      "loss": 0.1222,
+      "step": 20840
+    },
+    {
+      "epoch": 18.7,
+      "learning_rate": 3.475962497506483e-05,
+      "loss": 0.1031,
+      "step": 20850
+    },
+    {
+      "epoch": 18.71,
+      "learning_rate": 3.4749650907640134e-05,
+      "loss": 0.0738,
+      "step": 20860
+    },
+    {
+      "epoch": 18.72,
+      "learning_rate": 3.473967684021544e-05,
+      "loss": 0.0693,
+      "step": 20870
+    },
+    {
+      "epoch": 18.73,
+      "learning_rate": 3.4729702772790746e-05,
+      "loss": 0.1145,
+      "step": 20880
+    },
+    {
+      "epoch": 18.74,
+      "learning_rate": 3.471972870536605e-05,
+      "loss": 0.0771,
+      "step": 20890
+    },
+    {
+      "epoch": 18.75,
+      "learning_rate": 3.470975463794136e-05,
+      "loss": 0.0853,
+      "step": 20900
+    },
+    {
+      "epoch": 18.76,
+      "learning_rate": 3.469978057051666e-05,
+      "loss": 0.0942,
+      "step": 20910
+    },
+    {
+      "epoch": 18.77,
+      "learning_rate": 3.4689806503091964e-05,
+      "loss": 0.1007,
+      "step": 20920
+    },
+    {
+      "epoch": 18.78,
+      "learning_rate": 3.467983243566726e-05,
+      "loss": 0.0687,
+      "step": 20930
+    },
+    {
+      "epoch": 18.78,
+      "learning_rate": 3.466985836824257e-05,
+      "loss": 0.1179,
+      "step": 20940
+    },
+    {
+      "epoch": 18.79,
+      "learning_rate": 3.4659884300817876e-05,
+      "loss": 0.0788,
+      "step": 20950
+    },
+    {
+      "epoch": 18.8,
+      "learning_rate": 3.464991023339318e-05,
+      "loss": 0.0966,
+      "step": 20960
+    },
+    {
+      "epoch": 18.81,
+      "learning_rate": 3.463993616596848e-05,
+      "loss": 0.1133,
+      "step": 20970
+    },
+    {
+      "epoch": 18.82,
+      "learning_rate": 3.462996209854379e-05,
+      "loss": 0.0671,
+      "step": 20980
+    },
+    {
+      "epoch": 18.83,
+      "learning_rate": 3.4619988031119094e-05,
+      "loss": 0.1021,
+      "step": 20990
+    },
+    {
+      "epoch": 18.84,
+      "learning_rate": 3.46100139636944e-05,
+      "loss": 0.0872,
+      "step": 21000
+    },
+    {
+      "epoch": 18.85,
+      "learning_rate": 3.46000398962697e-05,
+      "loss": 0.0812,
+      "step": 21010
+    },
+    {
+      "epoch": 18.86,
+      "learning_rate": 3.4590065828845006e-05,
+      "loss": 0.1317,
+      "step": 21020
+    },
+    {
+      "epoch": 18.87,
+      "learning_rate": 3.4580091761420305e-05,
+      "loss": 0.0935,
+      "step": 21030
+    },
+    {
+      "epoch": 18.87,
+      "learning_rate": 3.457011769399561e-05,
+      "loss": 0.0717,
+      "step": 21040
+    },
+    {
+      "epoch": 18.88,
+      "learning_rate": 3.456014362657092e-05,
+      "loss": 0.0717,
+      "step": 21050
+    },
+    {
+      "epoch": 18.89,
+      "learning_rate": 3.4550169559146224e-05,
+      "loss": 0.1227,
+      "step": 21060
+    },
+    {
+      "epoch": 18.9,
+      "learning_rate": 3.454019549172153e-05,
+      "loss": 0.1476,
+      "step": 21070
+    },
+    {
+      "epoch": 18.91,
+      "learning_rate": 3.453022142429683e-05,
+      "loss": 0.1153,
+      "step": 21080
+    },
+    {
+      "epoch": 18.92,
+      "learning_rate": 3.452024735687213e-05,
+      "loss": 0.1076,
+      "step": 21090
+    },
+    {
+      "epoch": 18.93,
+      "learning_rate": 3.4510273289447435e-05,
+      "loss": 0.0735,
+      "step": 21100
+    },
+    {
+      "epoch": 18.94,
+      "learning_rate": 3.450029922202274e-05,
+      "loss": 0.0749,
+      "step": 21110
+    },
+    {
+      "epoch": 18.95,
+      "learning_rate": 3.449032515459805e-05,
+      "loss": 0.0719,
+      "step": 21120
+    },
+    {
+      "epoch": 18.95,
+      "learning_rate": 3.448035108717335e-05,
+      "loss": 0.0987,
+      "step": 21130
+    },
+    {
+      "epoch": 18.96,
+      "learning_rate": 3.447037701974866e-05,
+      "loss": 0.0848,
+      "step": 21140
+    },
+    {
+      "epoch": 18.97,
+      "learning_rate": 3.446040295232396e-05,
+      "loss": 0.0716,
+      "step": 21150
+    },
+    {
+      "epoch": 18.98,
+      "learning_rate": 3.4450428884899265e-05,
+      "loss": 0.0574,
+      "step": 21160
+    },
+    {
+      "epoch": 18.99,
+      "learning_rate": 3.4440454817474565e-05,
+      "loss": 0.0914,
+      "step": 21170
+    },
+    {
+      "epoch": 19.0,
+      "learning_rate": 3.443048075004987e-05,
+      "loss": 0.0839,
+      "step": 21180
+    },
+    {
+      "epoch": 19.0,
+      "eval_accuracy": {
+        "accuracy": 0.9770127831352321
+      },
+      "eval_f1": {
+        "f1": 0.9758154254345139
+      },
+      "eval_loss": 0.06907211244106293,
+      "eval_precision": {
+        "precision": 0.9761822643915987
+      },
+      "eval_recall": {
+        "recall": 0.9754580155091578
+      },
+      "eval_runtime": 197.4382,
+      "eval_samples_per_second": 90.337,
+      "eval_steps_per_second": 5.647,
+      "step": 21180
+    },
+    {
+      "epoch": 19.01,
+      "learning_rate": 3.442050668262518e-05,
+      "loss": 0.0947,
+      "step": 21190
+    },
+    {
+      "epoch": 19.02,
+      "learning_rate": 3.441053261520048e-05,
+      "loss": 0.063,
+      "step": 21200
+    },
+    {
+      "epoch": 19.03,
+      "learning_rate": 3.440055854777578e-05,
+      "loss": 0.0879,
+      "step": 21210
+    },
+    {
+      "epoch": 19.04,
+      "learning_rate": 3.439058448035109e-05,
+      "loss": 0.0552,
+      "step": 21220
+    },
+    {
+      "epoch": 19.04,
+      "learning_rate": 3.4380610412926395e-05,
+      "loss": 0.0411,
+      "step": 21230
+    },
+    {
+      "epoch": 19.05,
+      "learning_rate": 3.43706363455017e-05,
+      "loss": 0.0852,
+      "step": 21240
+    },
+    {
+      "epoch": 19.06,
+      "learning_rate": 3.4360662278077e-05,
+      "loss": 0.0891,
+      "step": 21250
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 3.435068821065231e-05,
+      "loss": 0.0618,
+      "step": 21260
+    },
+    {
+      "epoch": 19.08,
+      "learning_rate": 3.4340714143227606e-05,
+      "loss": 0.078,
+      "step": 21270
+    },
+    {
+      "epoch": 19.09,
+      "learning_rate": 3.433074007580291e-05,
+      "loss": 0.0723,
+      "step": 21280
+    },
+    {
+      "epoch": 19.1,
+      "learning_rate": 3.432076600837822e-05,
+      "loss": 0.0807,
+      "step": 21290
+    },
+    {
+      "epoch": 19.11,
+      "learning_rate": 3.4310791940953525e-05,
+      "loss": 0.114,
+      "step": 21300
+    },
+    {
+      "epoch": 19.12,
+      "learning_rate": 3.430081787352883e-05,
+      "loss": 0.0578,
+      "step": 21310
+    },
+    {
+      "epoch": 19.13,
+      "learning_rate": 3.429084380610413e-05,
+      "loss": 0.0716,
+      "step": 21320
+    },
+    {
+      "epoch": 19.13,
+      "learning_rate": 3.4280869738679436e-05,
+      "loss": 0.0806,
+      "step": 21330
+    },
+    {
+      "epoch": 19.14,
+      "learning_rate": 3.4270895671254736e-05,
+      "loss": 0.0926,
+      "step": 21340
+    },
+    {
+      "epoch": 19.15,
+      "learning_rate": 3.426092160383004e-05,
+      "loss": 0.0694,
+      "step": 21350
+    },
+    {
+      "epoch": 19.16,
+      "learning_rate": 3.425094753640535e-05,
+      "loss": 0.0628,
+      "step": 21360
+    },
+    {
+      "epoch": 19.17,
+      "learning_rate": 3.4240973468980654e-05,
+      "loss": 0.0757,
+      "step": 21370
+    },
+    {
+      "epoch": 19.18,
+      "learning_rate": 3.423099940155596e-05,
+      "loss": 0.1024,
+      "step": 21380
+    },
+    {
+      "epoch": 19.19,
+      "learning_rate": 3.422102533413126e-05,
+      "loss": 0.1002,
+      "step": 21390
+    },
+    {
+      "epoch": 19.2,
+      "learning_rate": 3.421105126670656e-05,
+      "loss": 0.1286,
+      "step": 21400
+    },
+    {
+      "epoch": 19.21,
+      "learning_rate": 3.4201077199281866e-05,
+      "loss": 0.0689,
+      "step": 21410
+    },
+    {
+      "epoch": 19.22,
+      "learning_rate": 3.419110313185717e-05,
+      "loss": 0.0944,
+      "step": 21420
+    },
+    {
+      "epoch": 19.22,
+      "learning_rate": 3.418112906443248e-05,
+      "loss": 0.0911,
+      "step": 21430
+    },
+    {
+      "epoch": 19.23,
+      "learning_rate": 3.4171154997007784e-05,
+      "loss": 0.0951,
+      "step": 21440
+    },
+    {
+      "epoch": 19.24,
+      "learning_rate": 3.4161180929583084e-05,
+      "loss": 0.0735,
+      "step": 21450
+    },
+    {
+      "epoch": 19.25,
+      "learning_rate": 3.415120686215839e-05,
+      "loss": 0.0771,
+      "step": 21460
+    },
+    {
+      "epoch": 19.26,
+      "learning_rate": 3.4141232794733696e-05,
+      "loss": 0.0406,
+      "step": 21470
+    },
+    {
+      "epoch": 19.27,
+      "learning_rate": 3.4131258727308995e-05,
+      "loss": 0.0741,
+      "step": 21480
+    },
+    {
+      "epoch": 19.28,
+      "learning_rate": 3.41212846598843e-05,
+      "loss": 0.1092,
+      "step": 21490
+    },
+    {
+      "epoch": 19.29,
+      "learning_rate": 3.411131059245961e-05,
+      "loss": 0.0806,
+      "step": 21500
+    },
+    {
+      "epoch": 19.3,
+      "learning_rate": 3.410133652503491e-05,
+      "loss": 0.0541,
+      "step": 21510
+    },
+    {
+      "epoch": 19.3,
+      "learning_rate": 3.4091362457610213e-05,
+      "loss": 0.0883,
+      "step": 21520
+    },
+    {
+      "epoch": 19.31,
+      "learning_rate": 3.408138839018552e-05,
+      "loss": 0.0668,
+      "step": 21530
+    },
+    {
+      "epoch": 19.32,
+      "learning_rate": 3.4071414322760826e-05,
+      "loss": 0.0774,
+      "step": 21540
+    },
+    {
+      "epoch": 19.33,
+      "learning_rate": 3.406144025533613e-05,
+      "loss": 0.0899,
+      "step": 21550
+    },
+    {
+      "epoch": 19.34,
+      "learning_rate": 3.405146618791143e-05,
+      "loss": 0.069,
+      "step": 21560
+    },
+    {
+      "epoch": 19.35,
+      "learning_rate": 3.404149212048674e-05,
+      "loss": 0.0988,
+      "step": 21570
+    },
+    {
+      "epoch": 19.36,
+      "learning_rate": 3.403151805306204e-05,
+      "loss": 0.0583,
+      "step": 21580
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.402154398563734e-05,
+      "loss": 0.0682,
+      "step": 21590
+    },
+    {
+      "epoch": 19.38,
+      "learning_rate": 3.401156991821265e-05,
+      "loss": 0.112,
+      "step": 21600
+    },
+    {
+      "epoch": 19.39,
+      "learning_rate": 3.4001595850787956e-05,
+      "loss": 0.0711,
+      "step": 21610
+    },
+    {
+      "epoch": 19.39,
+      "learning_rate": 3.399162178336326e-05,
+      "loss": 0.0903,
+      "step": 21620
+    },
+    {
+      "epoch": 19.4,
+      "learning_rate": 3.398164771593856e-05,
+      "loss": 0.0759,
+      "step": 21630
+    },
+    {
+      "epoch": 19.41,
+      "learning_rate": 3.397167364851386e-05,
+      "loss": 0.0535,
+      "step": 21640
+    },
+    {
+      "epoch": 19.42,
+      "learning_rate": 3.396169958108917e-05,
+      "loss": 0.1086,
+      "step": 21650
+    },
+    {
+      "epoch": 19.43,
+      "learning_rate": 3.395172551366447e-05,
+      "loss": 0.0778,
+      "step": 21660
+    },
+    {
+      "epoch": 19.44,
+      "learning_rate": 3.394175144623978e-05,
+      "loss": 0.0944,
+      "step": 21670
+    },
+    {
+      "epoch": 19.45,
+      "learning_rate": 3.3931777378815085e-05,
+      "loss": 0.0606,
+      "step": 21680
+    },
+    {
+      "epoch": 19.46,
+      "learning_rate": 3.3921803311390385e-05,
+      "loss": 0.0577,
+      "step": 21690
+    },
+    {
+      "epoch": 19.47,
+      "learning_rate": 3.391182924396569e-05,
+      "loss": 0.1027,
+      "step": 21700
+    },
+    {
+      "epoch": 19.48,
+      "learning_rate": 3.3901855176541e-05,
+      "loss": 0.0818,
+      "step": 21710
+    },
+    {
+      "epoch": 19.48,
+      "learning_rate": 3.38918811091163e-05,
+      "loss": 0.05,
+      "step": 21720
+    },
+    {
+      "epoch": 19.49,
+      "learning_rate": 3.38819070416916e-05,
+      "loss": 0.0815,
+      "step": 21730
+    },
+    {
+      "epoch": 19.5,
+      "learning_rate": 3.387193297426691e-05,
+      "loss": 0.0904,
+      "step": 21740
+    },
+    {
+      "epoch": 19.51,
+      "learning_rate": 3.386195890684221e-05,
+      "loss": 0.0933,
+      "step": 21750
+    },
+    {
+      "epoch": 19.52,
+      "learning_rate": 3.3851984839417515e-05,
+      "loss": 0.0868,
+      "step": 21760
+    },
+    {
+      "epoch": 19.53,
+      "learning_rate": 3.384201077199282e-05,
+      "loss": 0.0885,
+      "step": 21770
+    },
+    {
+      "epoch": 19.54,
+      "learning_rate": 3.383203670456813e-05,
+      "loss": 0.0719,
+      "step": 21780
+    },
+    {
+      "epoch": 19.55,
+      "learning_rate": 3.382206263714343e-05,
+      "loss": 0.0591,
+      "step": 21790
+    },
+    {
+      "epoch": 19.56,
+      "learning_rate": 3.381208856971873e-05,
+      "loss": 0.0673,
+      "step": 21800
+    },
+    {
+      "epoch": 19.56,
+      "learning_rate": 3.380211450229404e-05,
+      "loss": 0.0972,
+      "step": 21810
+    },
+    {
+      "epoch": 19.57,
+      "learning_rate": 3.379214043486934e-05,
+      "loss": 0.0604,
+      "step": 21820
+    },
+    {
+      "epoch": 19.58,
+      "learning_rate": 3.3782166367444644e-05,
+      "loss": 0.0987,
+      "step": 21830
+    },
+    {
+      "epoch": 19.59,
+      "learning_rate": 3.377219230001995e-05,
+      "loss": 0.0838,
+      "step": 21840
+    },
+    {
+      "epoch": 19.6,
+      "learning_rate": 3.376221823259526e-05,
+      "loss": 0.0705,
+      "step": 21850
+    },
+    {
+      "epoch": 19.61,
+      "learning_rate": 3.375224416517056e-05,
+      "loss": 0.074,
+      "step": 21860
+    },
+    {
+      "epoch": 19.62,
+      "learning_rate": 3.374227009774586e-05,
+      "loss": 0.097,
+      "step": 21870
+    },
+    {
+      "epoch": 19.63,
+      "learning_rate": 3.373229603032116e-05,
+      "loss": 0.0976,
+      "step": 21880
+    },
+    {
+      "epoch": 19.64,
+      "learning_rate": 3.372232196289647e-05,
+      "loss": 0.0693,
+      "step": 21890
+    },
+    {
+      "epoch": 19.65,
+      "learning_rate": 3.3712347895471774e-05,
+      "loss": 0.067,
+      "step": 21900
+    },
+    {
+      "epoch": 19.65,
+      "learning_rate": 3.370237382804708e-05,
+      "loss": 0.0594,
+      "step": 21910
+    },
+    {
+      "epoch": 19.66,
+      "learning_rate": 3.3692399760622387e-05,
+      "loss": 0.0837,
+      "step": 21920
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 3.3682425693197686e-05,
+      "loss": 0.1103,
+      "step": 21930
+    },
+    {
+      "epoch": 19.68,
+      "learning_rate": 3.367245162577299e-05,
+      "loss": 0.0871,
+      "step": 21940
+    },
+    {
+      "epoch": 19.69,
+      "learning_rate": 3.36624775583483e-05,
+      "loss": 0.0635,
+      "step": 21950
+    },
+    {
+      "epoch": 19.7,
+      "learning_rate": 3.36525034909236e-05,
+      "loss": 0.0653,
+      "step": 21960
+    },
+    {
+      "epoch": 19.71,
+      "learning_rate": 3.3642529423498904e-05,
+      "loss": 0.0869,
+      "step": 21970
+    },
+    {
+      "epoch": 19.72,
+      "learning_rate": 3.363255535607421e-05,
+      "loss": 0.0988,
+      "step": 21980
+    },
+    {
+      "epoch": 19.73,
+      "learning_rate": 3.362258128864951e-05,
+      "loss": 0.0642,
+      "step": 21990
+    },
+    {
+      "epoch": 19.74,
+      "learning_rate": 3.3612607221224816e-05,
+      "loss": 0.0657,
+      "step": 22000
+    },
+    {
+      "epoch": 19.74,
+      "learning_rate": 3.360263315380012e-05,
+      "loss": 0.0916,
+      "step": 22010
+    },
+    {
+      "epoch": 19.75,
+      "learning_rate": 3.359265908637543e-05,
+      "loss": 0.1009,
+      "step": 22020
+    },
+    {
+      "epoch": 19.76,
+      "learning_rate": 3.3582685018950734e-05,
+      "loss": 0.0612,
+      "step": 22030
+    },
+    {
+      "epoch": 19.77,
+      "learning_rate": 3.3572710951526034e-05,
+      "loss": 0.0846,
+      "step": 22040
+    },
+    {
+      "epoch": 19.78,
+      "learning_rate": 3.356273688410134e-05,
+      "loss": 0.0647,
+      "step": 22050
+    },
+    {
+      "epoch": 19.79,
+      "learning_rate": 3.355276281667664e-05,
+      "loss": 0.0744,
+      "step": 22060
+    },
+    {
+      "epoch": 19.8,
+      "learning_rate": 3.3542788749251946e-05,
+      "loss": 0.0889,
+      "step": 22070
+    },
+    {
+      "epoch": 19.81,
+      "learning_rate": 3.353281468182725e-05,
+      "loss": 0.1066,
+      "step": 22080
+    },
+    {
+      "epoch": 19.82,
+      "learning_rate": 3.352284061440256e-05,
+      "loss": 0.0595,
+      "step": 22090
+    },
+    {
+      "epoch": 19.83,
+      "learning_rate": 3.3512866546977864e-05,
+      "loss": 0.0899,
+      "step": 22100
+    },
+    {
+      "epoch": 19.83,
+      "learning_rate": 3.3502892479553164e-05,
+      "loss": 0.0708,
+      "step": 22110
+    },
+    {
+      "epoch": 19.84,
+      "learning_rate": 3.349291841212846e-05,
+      "loss": 0.0613,
+      "step": 22120
+    },
+    {
+      "epoch": 19.85,
+      "learning_rate": 3.348294434470377e-05,
+      "loss": 0.0742,
+      "step": 22130
+    },
+    {
+      "epoch": 19.86,
+      "learning_rate": 3.3472970277279075e-05,
+      "loss": 0.0982,
+      "step": 22140
+    },
+    {
+      "epoch": 19.87,
+      "learning_rate": 3.346299620985438e-05,
+      "loss": 0.0739,
+      "step": 22150
+    },
+    {
+      "epoch": 19.88,
+      "learning_rate": 3.345302214242969e-05,
+      "loss": 0.1002,
+      "step": 22160
+    },
+    {
+      "epoch": 19.89,
+      "learning_rate": 3.344304807500499e-05,
+      "loss": 0.0765,
+      "step": 22170
+    },
+    {
+      "epoch": 19.9,
+      "learning_rate": 3.343307400758029e-05,
+      "loss": 0.0932,
+      "step": 22180
+    },
+    {
+      "epoch": 19.91,
+      "learning_rate": 3.342309994015559e-05,
+      "loss": 0.0875,
+      "step": 22190
+    },
+    {
+      "epoch": 19.91,
+      "learning_rate": 3.34131258727309e-05,
+      "loss": 0.0788,
+      "step": 22200
+    },
+    {
+      "epoch": 19.92,
+      "learning_rate": 3.3403151805306205e-05,
+      "loss": 0.0727,
+      "step": 22210
+    },
+    {
+      "epoch": 19.93,
+      "learning_rate": 3.339317773788151e-05,
+      "loss": 0.0741,
+      "step": 22220
+    },
+    {
+      "epoch": 19.94,
+      "learning_rate": 3.338320367045681e-05,
+      "loss": 0.0752,
+      "step": 22230
+    },
+    {
+      "epoch": 19.95,
+      "learning_rate": 3.337322960303212e-05,
+      "loss": 0.0954,
+      "step": 22240
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 3.336325553560742e-05,
+      "loss": 0.0559,
+      "step": 22250
+    },
+    {
+      "epoch": 19.97,
+      "learning_rate": 3.335328146818273e-05,
+      "loss": 0.0666,
+      "step": 22260
+    },
+    {
+      "epoch": 19.98,
+      "learning_rate": 3.334330740075803e-05,
+      "loss": 0.0678,
+      "step": 22270
+    },
+    {
+      "epoch": 19.99,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.0851,
+      "step": 22280
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 3.332335926590864e-05,
+      "loss": 0.0752,
+      "step": 22290
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": {
+        "accuracy": 0.9765642520744562
+      },
+      "eval_f1": {
+        "f1": 0.9753527993174993
+      },
+      "eval_loss": 0.07372234016656876,
+      "eval_precision": {
+        "precision": 0.9756389600577889
+      },
+      "eval_recall": {
+        "recall": 0.9751662665128649
+      },
+      "eval_runtime": 185.001,
+      "eval_samples_per_second": 96.41,
+      "eval_steps_per_second": 6.027,
+      "step": 22295
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 3.331338519848394e-05,
+      "loss": 0.0697,
+      "step": 22300
+    },
+    {
+      "epoch": 20.01,
+      "learning_rate": 3.330341113105925e-05,
+      "loss": 0.0854,
+      "step": 22310
+    },
+    {
+      "epoch": 20.02,
+      "learning_rate": 3.329343706363455e-05,
+      "loss": 0.0683,
+      "step": 22320
+    },
+    {
+      "epoch": 20.03,
+      "learning_rate": 3.328346299620986e-05,
+      "loss": 0.0593,
+      "step": 22330
+    },
+    {
+      "epoch": 20.04,
+      "learning_rate": 3.3273488928785165e-05,
+      "loss": 0.074,
+      "step": 22340
+    },
+    {
+      "epoch": 20.05,
+      "learning_rate": 3.3263514861360465e-05,
+      "loss": 0.0987,
+      "step": 22350
+    },
+    {
+      "epoch": 20.06,
+      "learning_rate": 3.3253540793935764e-05,
+      "loss": 0.0949,
+      "step": 22360
+    },
+    {
+      "epoch": 20.07,
+      "learning_rate": 3.324356672651107e-05,
+      "loss": 0.0582,
+      "step": 22370
+    },
+    {
+      "epoch": 20.08,
+      "learning_rate": 3.3233592659086377e-05,
+      "loss": 0.0503,
+      "step": 22380
+    },
+    {
+      "epoch": 20.09,
+      "learning_rate": 3.322361859166168e-05,
+      "loss": 0.0682,
+      "step": 22390
+    },
+    {
+      "epoch": 20.09,
+      "learning_rate": 3.321364452423699e-05,
+      "loss": 0.0886,
+      "step": 22400
+    },
+    {
+      "epoch": 20.1,
+      "learning_rate": 3.320367045681229e-05,
+      "loss": 0.0756,
+      "step": 22410
+    },
+    {
+      "epoch": 20.11,
+      "learning_rate": 3.3193696389387595e-05,
+      "loss": 0.0461,
+      "step": 22420
+    },
+    {
+      "epoch": 20.12,
+      "learning_rate": 3.3183722321962894e-05,
+      "loss": 0.093,
+      "step": 22430
+    },
+    {
+      "epoch": 20.13,
+      "learning_rate": 3.31737482545382e-05,
+      "loss": 0.1076,
+      "step": 22440
+    },
+    {
+      "epoch": 20.14,
+      "learning_rate": 3.3163774187113506e-05,
+      "loss": 0.0715,
+      "step": 22450
+    },
+    {
+      "epoch": 20.15,
+      "learning_rate": 3.315380011968881e-05,
+      "loss": 0.0941,
+      "step": 22460
+    },
+    {
+      "epoch": 20.16,
+      "learning_rate": 3.314382605226411e-05,
+      "loss": 0.0662,
+      "step": 22470
+    },
+    {
+      "epoch": 20.17,
+      "learning_rate": 3.313385198483942e-05,
+      "loss": 0.063,
+      "step": 22480
+    },
+    {
+      "epoch": 20.17,
+      "learning_rate": 3.3123877917414724e-05,
+      "loss": 0.0682,
+      "step": 22490
+    },
+    {
+      "epoch": 20.18,
+      "learning_rate": 3.311390384999003e-05,
+      "loss": 0.067,
+      "step": 22500
+    },
+    {
+      "epoch": 20.19,
+      "learning_rate": 3.310392978256533e-05,
+      "loss": 0.0847,
+      "step": 22510
+    },
+    {
+      "epoch": 20.2,
+      "learning_rate": 3.3093955715140636e-05,
+      "loss": 0.0859,
+      "step": 22520
+    },
+    {
+      "epoch": 20.21,
+      "learning_rate": 3.308398164771594e-05,
+      "loss": 0.0892,
+      "step": 22530
+    },
+    {
+      "epoch": 20.22,
+      "learning_rate": 3.307400758029124e-05,
+      "loss": 0.0671,
+      "step": 22540
+    },
+    {
+      "epoch": 20.23,
+      "learning_rate": 3.306403351286655e-05,
+      "loss": 0.0645,
+      "step": 22550
+    },
+    {
+      "epoch": 20.24,
+      "learning_rate": 3.3054059445441854e-05,
+      "loss": 0.093,
+      "step": 22560
+    },
+    {
+      "epoch": 20.25,
+      "learning_rate": 3.304408537801716e-05,
+      "loss": 0.1148,
+      "step": 22570
+    },
+    {
+      "epoch": 20.26,
+      "learning_rate": 3.3034111310592466e-05,
+      "loss": 0.0747,
+      "step": 22580
+    },
+    {
+      "epoch": 20.26,
+      "learning_rate": 3.3024137243167766e-05,
+      "loss": 0.0829,
+      "step": 22590
+    },
+    {
+      "epoch": 20.27,
+      "learning_rate": 3.3014163175743065e-05,
+      "loss": 0.0829,
+      "step": 22600
+    },
+    {
+      "epoch": 20.28,
+      "learning_rate": 3.300418910831837e-05,
+      "loss": 0.0722,
+      "step": 22610
+    },
+    {
+      "epoch": 20.29,
+      "learning_rate": 3.299421504089368e-05,
+      "loss": 0.0785,
+      "step": 22620
+    },
+    {
+      "epoch": 20.3,
+      "learning_rate": 3.2984240973468984e-05,
+      "loss": 0.0887,
+      "step": 22630
+    },
+    {
+      "epoch": 20.31,
+      "learning_rate": 3.297426690604429e-05,
+      "loss": 0.0803,
+      "step": 22640
+    },
+    {
+      "epoch": 20.32,
+      "learning_rate": 3.296429283861959e-05,
+      "loss": 0.0906,
+      "step": 22650
+    },
+    {
+      "epoch": 20.33,
+      "learning_rate": 3.2954318771194896e-05,
+      "loss": 0.0919,
+      "step": 22660
+    },
+    {
+      "epoch": 20.34,
+      "learning_rate": 3.2944344703770195e-05,
+      "loss": 0.0727,
+      "step": 22670
+    },
+    {
+      "epoch": 20.35,
+      "learning_rate": 3.29343706363455e-05,
+      "loss": 0.0811,
+      "step": 22680
+    },
+    {
+      "epoch": 20.35,
+      "learning_rate": 3.292439656892081e-05,
+      "loss": 0.0726,
+      "step": 22690
+    },
+    {
+      "epoch": 20.36,
+      "learning_rate": 3.2914422501496114e-05,
+      "loss": 0.0641,
+      "step": 22700
+    },
+    {
+      "epoch": 20.37,
+      "learning_rate": 3.290444843407142e-05,
+      "loss": 0.0777,
+      "step": 22710
+    },
+    {
+      "epoch": 20.38,
+      "learning_rate": 3.289447436664672e-05,
+      "loss": 0.0513,
+      "step": 22720
+    },
+    {
+      "epoch": 20.39,
+      "learning_rate": 3.2884500299222025e-05,
+      "loss": 0.0857,
+      "step": 22730
+    },
+    {
+      "epoch": 20.4,
+      "learning_rate": 3.287452623179733e-05,
+      "loss": 0.1225,
+      "step": 22740
+    },
+    {
+      "epoch": 20.41,
+      "learning_rate": 3.286455216437263e-05,
+      "loss": 0.1103,
+      "step": 22750
+    },
+    {
+      "epoch": 20.42,
+      "learning_rate": 3.285457809694794e-05,
+      "loss": 0.0675,
+      "step": 22760
+    },
+    {
+      "epoch": 20.43,
+      "learning_rate": 3.2844604029523243e-05,
+      "loss": 0.0853,
+      "step": 22770
+    },
+    {
+      "epoch": 20.44,
+      "learning_rate": 3.283462996209854e-05,
+      "loss": 0.0871,
+      "step": 22780
+    },
+    {
+      "epoch": 20.44,
+      "learning_rate": 3.282465589467385e-05,
+      "loss": 0.086,
+      "step": 22790
+    },
+    {
+      "epoch": 20.45,
+      "learning_rate": 3.2814681827249155e-05,
+      "loss": 0.086,
+      "step": 22800
+    },
+    {
+      "epoch": 20.46,
+      "learning_rate": 3.280470775982446e-05,
+      "loss": 0.0802,
+      "step": 22810
+    },
+    {
+      "epoch": 20.47,
+      "learning_rate": 3.279473369239976e-05,
+      "loss": 0.0735,
+      "step": 22820
+    },
+    {
+      "epoch": 20.48,
+      "learning_rate": 3.278475962497507e-05,
+      "loss": 0.1059,
+      "step": 22830
+    },
+    {
+      "epoch": 20.49,
+      "learning_rate": 3.2774785557550366e-05,
+      "loss": 0.0969,
+      "step": 22840
+    },
+    {
+      "epoch": 20.5,
+      "learning_rate": 3.276481149012567e-05,
+      "loss": 0.0706,
+      "step": 22850
+    },
+    {
+      "epoch": 20.51,
+      "learning_rate": 3.275483742270098e-05,
+      "loss": 0.0902,
+      "step": 22860
+    },
+    {
+      "epoch": 20.52,
+      "learning_rate": 3.2744863355276285e-05,
+      "loss": 0.0633,
+      "step": 22870
+    },
+    {
+      "epoch": 20.52,
+      "learning_rate": 3.273488928785159e-05,
+      "loss": 0.082,
+      "step": 22880
+    },
+    {
+      "epoch": 20.53,
+      "learning_rate": 3.272491522042689e-05,
+      "loss": 0.0533,
+      "step": 22890
+    },
+    {
+      "epoch": 20.54,
+      "learning_rate": 3.271494115300219e-05,
+      "loss": 0.0936,
+      "step": 22900
+    },
+    {
+      "epoch": 20.55,
+      "learning_rate": 3.2704967085577496e-05,
+      "loss": 0.0736,
+      "step": 22910
+    },
+    {
+      "epoch": 20.56,
+      "learning_rate": 3.26949930181528e-05,
+      "loss": 0.1105,
+      "step": 22920
+    },
+    {
+      "epoch": 20.57,
+      "learning_rate": 3.268501895072811e-05,
+      "loss": 0.0874,
+      "step": 22930
+    },
+    {
+      "epoch": 20.58,
+      "learning_rate": 3.2675044883303415e-05,
+      "loss": 0.061,
+      "step": 22940
+    },
+    {
+      "epoch": 20.59,
+      "learning_rate": 3.266507081587872e-05,
+      "loss": 0.0787,
+      "step": 22950
+    },
+    {
+      "epoch": 20.6,
+      "learning_rate": 3.265509674845402e-05,
+      "loss": 0.0714,
+      "step": 22960
+    },
+    {
+      "epoch": 20.61,
+      "learning_rate": 3.264512268102933e-05,
+      "loss": 0.0742,
+      "step": 22970
+    },
+    {
+      "epoch": 20.61,
+      "learning_rate": 3.2635148613604626e-05,
+      "loss": 0.0729,
+      "step": 22980
+    },
+    {
+      "epoch": 20.62,
+      "learning_rate": 3.262517454617993e-05,
+      "loss": 0.081,
+      "step": 22990
+    },
+    {
+      "epoch": 20.63,
+      "learning_rate": 3.261520047875524e-05,
+      "loss": 0.1077,
+      "step": 23000
+    },
+    {
+      "epoch": 20.64,
+      "learning_rate": 3.2605226411330545e-05,
+      "loss": 0.0967,
+      "step": 23010
+    },
+    {
+      "epoch": 20.65,
+      "learning_rate": 3.2595252343905844e-05,
+      "loss": 0.0874,
+      "step": 23020
+    },
+    {
+      "epoch": 20.66,
+      "learning_rate": 3.258527827648115e-05,
+      "loss": 0.0525,
+      "step": 23030
+    },
+    {
+      "epoch": 20.67,
+      "learning_rate": 3.2575304209056456e-05,
+      "loss": 0.085,
+      "step": 23040
+    },
+    {
+      "epoch": 20.68,
+      "learning_rate": 3.256533014163176e-05,
+      "loss": 0.0876,
+      "step": 23050
+    },
+    {
+      "epoch": 20.69,
+      "learning_rate": 3.255535607420706e-05,
+      "loss": 0.0731,
+      "step": 23060
+    },
+    {
+      "epoch": 20.7,
+      "learning_rate": 3.254538200678237e-05,
+      "loss": 0.0737,
+      "step": 23070
+    },
+    {
+      "epoch": 20.7,
+      "learning_rate": 3.253540793935767e-05,
+      "loss": 0.0848,
+      "step": 23080
+    },
+    {
+      "epoch": 20.71,
+      "learning_rate": 3.2525433871932974e-05,
+      "loss": 0.0795,
+      "step": 23090
+    },
+    {
+      "epoch": 20.72,
+      "learning_rate": 3.251545980450828e-05,
+      "loss": 0.067,
+      "step": 23100
+    },
+    {
+      "epoch": 20.73,
+      "learning_rate": 3.2505485737083586e-05,
+      "loss": 0.087,
+      "step": 23110
+    },
+    {
+      "epoch": 20.74,
+      "learning_rate": 3.249551166965889e-05,
+      "loss": 0.0811,
+      "step": 23120
+    },
+    {
+      "epoch": 20.75,
+      "learning_rate": 3.248553760223419e-05,
+      "loss": 0.089,
+      "step": 23130
+    },
+    {
+      "epoch": 20.76,
+      "learning_rate": 3.247556353480949e-05,
+      "loss": 0.0919,
+      "step": 23140
+    },
+    {
+      "epoch": 20.77,
+      "learning_rate": 3.24655894673848e-05,
+      "loss": 0.0522,
+      "step": 23150
+    },
+    {
+      "epoch": 20.78,
+      "learning_rate": 3.2455615399960104e-05,
+      "loss": 0.0723,
+      "step": 23160
+    },
+    {
+      "epoch": 20.78,
+      "learning_rate": 3.244564133253541e-05,
+      "loss": 0.0746,
+      "step": 23170
+    },
+    {
+      "epoch": 20.79,
+      "learning_rate": 3.2435667265110716e-05,
+      "loss": 0.0793,
+      "step": 23180
+    },
+    {
+      "epoch": 20.8,
+      "learning_rate": 3.242569319768602e-05,
+      "loss": 0.1083,
+      "step": 23190
+    },
+    {
+      "epoch": 20.81,
+      "learning_rate": 3.241571913026132e-05,
+      "loss": 0.1005,
+      "step": 23200
+    },
+    {
+      "epoch": 20.82,
+      "learning_rate": 3.240574506283663e-05,
+      "loss": 0.0805,
+      "step": 23210
+    },
+    {
+      "epoch": 20.83,
+      "learning_rate": 3.239577099541193e-05,
+      "loss": 0.079,
+      "step": 23220
+    },
+    {
+      "epoch": 20.84,
+      "learning_rate": 3.2385796927987233e-05,
+      "loss": 0.0919,
+      "step": 23230
+    },
+    {
+      "epoch": 20.85,
+      "learning_rate": 3.237582286056254e-05,
+      "loss": 0.0805,
+      "step": 23240
+    },
+    {
+      "epoch": 20.86,
+      "learning_rate": 3.2365848793137846e-05,
+      "loss": 0.0753,
+      "step": 23250
+    },
+    {
+      "epoch": 20.87,
+      "learning_rate": 3.2355874725713145e-05,
+      "loss": 0.0883,
+      "step": 23260
+    },
+    {
+      "epoch": 20.87,
+      "learning_rate": 3.234590065828845e-05,
+      "loss": 0.0877,
+      "step": 23270
+    },
+    {
+      "epoch": 20.88,
+      "learning_rate": 3.233592659086376e-05,
+      "loss": 0.0519,
+      "step": 23280
+    },
+    {
+      "epoch": 20.89,
+      "learning_rate": 3.2325952523439064e-05,
+      "loss": 0.0996,
+      "step": 23290
+    },
+    {
+      "epoch": 20.9,
+      "learning_rate": 3.231597845601436e-05,
+      "loss": 0.1142,
+      "step": 23300
+    },
+    {
+      "epoch": 20.91,
+      "learning_rate": 3.230600438858967e-05,
+      "loss": 0.0492,
+      "step": 23310
+    },
+    {
+      "epoch": 20.92,
+      "learning_rate": 3.229603032116497e-05,
+      "loss": 0.0903,
+      "step": 23320
+    },
+    {
+      "epoch": 20.93,
+      "learning_rate": 3.2286056253740275e-05,
+      "loss": 0.0741,
+      "step": 23330
+    },
+    {
+      "epoch": 20.94,
+      "learning_rate": 3.227608218631558e-05,
+      "loss": 0.0857,
+      "step": 23340
+    },
+    {
+      "epoch": 20.95,
+      "learning_rate": 3.226610811889089e-05,
+      "loss": 0.0474,
+      "step": 23350
+    },
+    {
+      "epoch": 20.96,
+      "learning_rate": 3.2256134051466194e-05,
+      "loss": 0.0605,
+      "step": 23360
+    },
+    {
+      "epoch": 20.96,
+      "learning_rate": 3.224615998404149e-05,
+      "loss": 0.0875,
+      "step": 23370
+    },
+    {
+      "epoch": 20.97,
+      "learning_rate": 3.223618591661679e-05,
+      "loss": 0.0698,
+      "step": 23380
+    },
+    {
+      "epoch": 20.98,
+      "learning_rate": 3.22262118491921e-05,
+      "loss": 0.0721,
+      "step": 23390
+    },
+    {
+      "epoch": 20.99,
+      "learning_rate": 3.2216237781767405e-05,
+      "loss": 0.0777,
+      "step": 23400
+    },
+    {
+      "epoch": 21.0,
+      "eval_accuracy": {
+        "accuracy": 0.9806010316214397
+      },
+      "eval_f1": {
+        "f1": 0.9797190165517261
+      },
+      "eval_loss": 0.05940975993871689,
+      "eval_precision": {
+        "precision": 0.9797694892510824
+      },
+      "eval_recall": {
+        "recall": 0.9796726937896714
+      },
+      "eval_runtime": 199.1046,
+      "eval_samples_per_second": 89.581,
+      "eval_steps_per_second": 5.6,
+      "step": 23409
+    },
+    {
+      "epoch": 21.0,
+      "learning_rate": 3.220626371434271e-05,
+      "loss": 0.079,
+      "step": 23410
+    },
+    {
+      "epoch": 21.01,
+      "learning_rate": 3.219628964691802e-05,
+      "loss": 0.072,
+      "step": 23420
+    },
+    {
+      "epoch": 21.02,
+      "learning_rate": 3.218631557949332e-05,
+      "loss": 0.0805,
+      "step": 23430
+    },
+    {
+      "epoch": 21.03,
+      "learning_rate": 3.217634151206862e-05,
+      "loss": 0.0415,
+      "step": 23440
+    },
+    {
+      "epoch": 21.04,
+      "learning_rate": 3.216636744464393e-05,
+      "loss": 0.1084,
+      "step": 23450
+    },
+    {
+      "epoch": 21.05,
+      "learning_rate": 3.215639337721923e-05,
+      "loss": 0.0675,
+      "step": 23460
+    },
+    {
+      "epoch": 21.05,
+      "learning_rate": 3.2146419309794535e-05,
+      "loss": 0.0643,
+      "step": 23470
+    },
+    {
+      "epoch": 21.06,
+      "learning_rate": 3.213644524236984e-05,
+      "loss": 0.0536,
+      "step": 23480
+    },
+    {
+      "epoch": 21.07,
+      "learning_rate": 3.212647117494515e-05,
+      "loss": 0.0528,
+      "step": 23490
+    },
+    {
+      "epoch": 21.08,
+      "learning_rate": 3.2116497107520446e-05,
+      "loss": 0.0374,
+      "step": 23500
+    },
+    {
+      "epoch": 21.09,
+      "learning_rate": 3.210652304009575e-05,
+      "loss": 0.0815,
+      "step": 23510
+    },
+    {
+      "epoch": 21.1,
+      "learning_rate": 3.209654897267106e-05,
+      "loss": 0.077,
+      "step": 23520
+    },
+    {
+      "epoch": 21.11,
+      "learning_rate": 3.208657490524636e-05,
+      "loss": 0.054,
+      "step": 23530
+    },
+    {
+      "epoch": 21.12,
+      "learning_rate": 3.2076600837821664e-05,
+      "loss": 0.1037,
+      "step": 23540
+    },
+    {
+      "epoch": 21.13,
+      "learning_rate": 3.206662677039697e-05,
+      "loss": 0.0581,
+      "step": 23550
+    },
+    {
+      "epoch": 21.13,
+      "learning_rate": 3.205665270297227e-05,
+      "loss": 0.0902,
+      "step": 23560
+    },
+    {
+      "epoch": 21.14,
+      "learning_rate": 3.2046678635547576e-05,
+      "loss": 0.0538,
+      "step": 23570
+    },
+    {
+      "epoch": 21.15,
+      "learning_rate": 3.203670456812288e-05,
+      "loss": 0.0928,
+      "step": 23580
+    },
+    {
+      "epoch": 21.16,
+      "learning_rate": 3.202673050069819e-05,
+      "loss": 0.0589,
+      "step": 23590
+    },
+    {
+      "epoch": 21.17,
+      "learning_rate": 3.2016756433273495e-05,
+      "loss": 0.0783,
+      "step": 23600
+    },
+    {
+      "epoch": 21.18,
+      "learning_rate": 3.2006782365848794e-05,
+      "loss": 0.0822,
+      "step": 23610
+    },
+    {
+      "epoch": 21.19,
+      "learning_rate": 3.1996808298424094e-05,
+      "loss": 0.0642,
+      "step": 23620
+    },
+    {
+      "epoch": 21.2,
+      "learning_rate": 3.19868342309994e-05,
+      "loss": 0.077,
+      "step": 23630
+    },
+    {
+      "epoch": 21.21,
+      "learning_rate": 3.1976860163574706e-05,
+      "loss": 0.0836,
+      "step": 23640
+    },
+    {
+      "epoch": 21.22,
+      "learning_rate": 3.196688609615001e-05,
+      "loss": 0.0886,
+      "step": 23650
+    },
+    {
+      "epoch": 21.22,
+      "learning_rate": 3.195691202872532e-05,
+      "loss": 0.0934,
+      "step": 23660
+    },
+    {
+      "epoch": 21.23,
+      "learning_rate": 3.1946937961300625e-05,
+      "loss": 0.0765,
+      "step": 23670
+    },
+    {
+      "epoch": 21.24,
+      "learning_rate": 3.1936963893875924e-05,
+      "loss": 0.0786,
+      "step": 23680
+    },
+    {
+      "epoch": 21.25,
+      "learning_rate": 3.192698982645122e-05,
+      "loss": 0.0757,
+      "step": 23690
+    },
+    {
+      "epoch": 21.26,
+      "learning_rate": 3.191701575902653e-05,
+      "loss": 0.0727,
+      "step": 23700
+    },
+    {
+      "epoch": 21.27,
+      "learning_rate": 3.1907041691601836e-05,
+      "loss": 0.0742,
+      "step": 23710
+    },
+    {
+      "epoch": 21.28,
+      "learning_rate": 3.189706762417714e-05,
+      "loss": 0.0737,
+      "step": 23720
+    },
+    {
+      "epoch": 21.29,
+      "learning_rate": 3.188709355675245e-05,
+      "loss": 0.0789,
+      "step": 23730
+    },
+    {
+      "epoch": 21.3,
+      "learning_rate": 3.187711948932775e-05,
+      "loss": 0.087,
+      "step": 23740
+    },
+    {
+      "epoch": 21.31,
+      "learning_rate": 3.1867145421903054e-05,
+      "loss": 0.0833,
+      "step": 23750
+    },
+    {
+      "epoch": 21.31,
+      "learning_rate": 3.185717135447836e-05,
+      "loss": 0.1103,
+      "step": 23760
+    },
+    {
+      "epoch": 21.32,
+      "learning_rate": 3.184719728705366e-05,
+      "loss": 0.0888,
+      "step": 23770
+    },
+    {
+      "epoch": 21.33,
+      "learning_rate": 3.1837223219628966e-05,
+      "loss": 0.076,
+      "step": 23780
+    },
+    {
+      "epoch": 21.34,
+      "learning_rate": 3.182724915220427e-05,
+      "loss": 0.0913,
+      "step": 23790
+    },
+    {
+      "epoch": 21.35,
+      "learning_rate": 3.181727508477957e-05,
+      "loss": 0.1204,
+      "step": 23800
+    },
+    {
+      "epoch": 21.36,
+      "learning_rate": 3.180730101735488e-05,
+      "loss": 0.0517,
+      "step": 23810
+    },
+    {
+      "epoch": 21.37,
+      "learning_rate": 3.1797326949930184e-05,
+      "loss": 0.05,
+      "step": 23820
+    },
+    {
+      "epoch": 21.38,
+      "learning_rate": 3.178735288250549e-05,
+      "loss": 0.0808,
+      "step": 23830
+    },
+    {
+      "epoch": 21.39,
+      "learning_rate": 3.1777378815080796e-05,
+      "loss": 0.0435,
+      "step": 23840
+    },
+    {
+      "epoch": 21.39,
+      "learning_rate": 3.1767404747656095e-05,
+      "loss": 0.0576,
+      "step": 23850
+    },
+    {
+      "epoch": 21.4,
+      "learning_rate": 3.17574306802314e-05,
+      "loss": 0.0588,
+      "step": 23860
+    },
+    {
+      "epoch": 21.41,
+      "learning_rate": 3.17474566128067e-05,
+      "loss": 0.0676,
+      "step": 23870
+    },
+    {
+      "epoch": 21.42,
+      "learning_rate": 3.173748254538201e-05,
+      "loss": 0.0652,
+      "step": 23880
+    },
+    {
+      "epoch": 21.43,
+      "learning_rate": 3.172750847795731e-05,
+      "loss": 0.0817,
+      "step": 23890
+    },
+    {
+      "epoch": 21.44,
+      "learning_rate": 3.171753441053262e-05,
+      "loss": 0.0692,
+      "step": 23900
+    },
+    {
+      "epoch": 21.45,
+      "learning_rate": 3.1707560343107926e-05,
+      "loss": 0.1341,
+      "step": 23910
+    },
+    {
+      "epoch": 21.46,
+      "learning_rate": 3.1697586275683225e-05,
+      "loss": 0.1038,
+      "step": 23920
+    },
+    {
+      "epoch": 21.47,
+      "learning_rate": 3.1687612208258525e-05,
+      "loss": 0.0797,
+      "step": 23930
+    },
+    {
+      "epoch": 21.48,
+      "learning_rate": 3.167763814083383e-05,
+      "loss": 0.1063,
+      "step": 23940
+    },
+    {
+      "epoch": 21.48,
+      "learning_rate": 3.166766407340914e-05,
+      "loss": 0.0503,
+      "step": 23950
+    },
+    {
+      "epoch": 21.49,
+      "learning_rate": 3.165769000598444e-05,
+      "loss": 0.0669,
+      "step": 23960
+    },
+    {
+      "epoch": 21.5,
+      "learning_rate": 3.164771593855975e-05,
+      "loss": 0.0844,
+      "step": 23970
+    },
+    {
+      "epoch": 21.51,
+      "learning_rate": 3.163774187113505e-05,
+      "loss": 0.0884,
+      "step": 23980
+    },
+    {
+      "epoch": 21.52,
+      "learning_rate": 3.1627767803710355e-05,
+      "loss": 0.0717,
+      "step": 23990
+    },
+    {
+      "epoch": 21.53,
+      "learning_rate": 3.161779373628566e-05,
+      "loss": 0.0667,
+      "step": 24000
+    },
+    {
+      "epoch": 21.54,
+      "learning_rate": 3.160781966886096e-05,
+      "loss": 0.0877,
+      "step": 24010
+    },
+    {
+      "epoch": 21.55,
+      "learning_rate": 3.159784560143627e-05,
+      "loss": 0.0516,
+      "step": 24020
+    },
+    {
+      "epoch": 21.56,
+      "learning_rate": 3.158787153401157e-05,
+      "loss": 0.0986,
+      "step": 24030
+    },
+    {
+      "epoch": 21.57,
+      "learning_rate": 3.157789746658687e-05,
+      "loss": 0.0585,
+      "step": 24040
+    },
+    {
+      "epoch": 21.57,
+      "learning_rate": 3.156792339916218e-05,
+      "loss": 0.0789,
+      "step": 24050
+    },
+    {
+      "epoch": 21.58,
+      "learning_rate": 3.1557949331737485e-05,
+      "loss": 0.0797,
+      "step": 24060
+    },
+    {
+      "epoch": 21.59,
+      "learning_rate": 3.154797526431279e-05,
+      "loss": 0.0838,
+      "step": 24070
+    },
+    {
+      "epoch": 21.6,
+      "learning_rate": 3.15380011968881e-05,
+      "loss": 0.0582,
+      "step": 24080
+    },
+    {
+      "epoch": 21.61,
+      "learning_rate": 3.1528027129463396e-05,
+      "loss": 0.0894,
+      "step": 24090
+    },
+    {
+      "epoch": 21.62,
+      "learning_rate": 3.15180530620387e-05,
+      "loss": 0.0778,
+      "step": 24100
+    },
+    {
+      "epoch": 21.63,
+      "learning_rate": 3.1508078994614e-05,
+      "loss": 0.0787,
+      "step": 24110
+    },
+    {
+      "epoch": 21.64,
+      "learning_rate": 3.149810492718931e-05,
+      "loss": 0.0805,
+      "step": 24120
+    },
+    {
+      "epoch": 21.65,
+      "learning_rate": 3.1488130859764614e-05,
+      "loss": 0.0551,
+      "step": 24130
+    },
+    {
+      "epoch": 21.66,
+      "learning_rate": 3.147815679233992e-05,
+      "loss": 0.1026,
+      "step": 24140
+    },
+    {
+      "epoch": 21.66,
+      "learning_rate": 3.146818272491523e-05,
+      "loss": 0.0847,
+      "step": 24150
+    },
+    {
+      "epoch": 21.67,
+      "learning_rate": 3.1458208657490526e-05,
+      "loss": 0.0546,
+      "step": 24160
+    },
+    {
+      "epoch": 21.68,
+      "learning_rate": 3.1448234590065826e-05,
+      "loss": 0.0901,
+      "step": 24170
+    },
+    {
+      "epoch": 21.69,
+      "learning_rate": 3.143826052264113e-05,
+      "loss": 0.0789,
+      "step": 24180
+    },
+    {
+      "epoch": 21.7,
+      "learning_rate": 3.142828645521644e-05,
+      "loss": 0.0921,
+      "step": 24190
+    },
+    {
+      "epoch": 21.71,
+      "learning_rate": 3.1418312387791744e-05,
+      "loss": 0.0926,
+      "step": 24200
+    },
+    {
+      "epoch": 21.72,
+      "learning_rate": 3.140833832036705e-05,
+      "loss": 0.0723,
+      "step": 24210
+    },
+    {
+      "epoch": 21.73,
+      "learning_rate": 3.139836425294235e-05,
+      "loss": 0.0861,
+      "step": 24220
+    },
+    {
+      "epoch": 21.74,
+      "learning_rate": 3.1388390185517656e-05,
+      "loss": 0.0614,
+      "step": 24230
+    },
+    {
+      "epoch": 21.74,
+      "learning_rate": 3.137841611809296e-05,
+      "loss": 0.1016,
+      "step": 24240
+    },
+    {
+      "epoch": 21.75,
+      "learning_rate": 3.136844205066826e-05,
+      "loss": 0.0838,
+      "step": 24250
+    },
+    {
+      "epoch": 21.76,
+      "learning_rate": 3.135846798324357e-05,
+      "loss": 0.0639,
+      "step": 24260
+    },
+    {
+      "epoch": 21.77,
+      "learning_rate": 3.1348493915818874e-05,
+      "loss": 0.0906,
+      "step": 24270
+    },
+    {
+      "epoch": 21.78,
+      "learning_rate": 3.1338519848394173e-05,
+      "loss": 0.062,
+      "step": 24280
+    },
+    {
+      "epoch": 21.79,
+      "learning_rate": 3.132854578096948e-05,
+      "loss": 0.0697,
+      "step": 24290
+    },
+    {
+      "epoch": 21.8,
+      "learning_rate": 3.1318571713544786e-05,
+      "loss": 0.0669,
+      "step": 24300
+    },
+    {
+      "epoch": 21.81,
+      "learning_rate": 3.130859764612009e-05,
+      "loss": 0.0533,
+      "step": 24310
+    },
+    {
+      "epoch": 21.82,
+      "learning_rate": 3.129862357869539e-05,
+      "loss": 0.0947,
+      "step": 24320
+    },
+    {
+      "epoch": 21.83,
+      "learning_rate": 3.12886495112707e-05,
+      "loss": 0.0936,
+      "step": 24330
+    },
+    {
+      "epoch": 21.83,
+      "learning_rate": 3.1278675443846004e-05,
+      "loss": 0.0537,
+      "step": 24340
+    },
+    {
+      "epoch": 21.84,
+      "learning_rate": 3.12687013764213e-05,
+      "loss": 0.078,
+      "step": 24350
+    },
+    {
+      "epoch": 21.85,
+      "learning_rate": 3.125872730899661e-05,
+      "loss": 0.0797,
+      "step": 24360
+    },
+    {
+      "epoch": 21.86,
+      "learning_rate": 3.1248753241571916e-05,
+      "loss": 0.0736,
+      "step": 24370
+    },
+    {
+      "epoch": 21.87,
+      "learning_rate": 3.123877917414722e-05,
+      "loss": 0.0574,
+      "step": 24380
+    },
+    {
+      "epoch": 21.88,
+      "learning_rate": 3.122880510672253e-05,
+      "loss": 0.0915,
+      "step": 24390
+    },
+    {
+      "epoch": 21.89,
+      "learning_rate": 3.121883103929783e-05,
+      "loss": 0.1133,
+      "step": 24400
+    },
+    {
+      "epoch": 21.9,
+      "learning_rate": 3.120885697187313e-05,
+      "loss": 0.0897,
+      "step": 24410
+    },
+    {
+      "epoch": 21.91,
+      "learning_rate": 3.119888290444843e-05,
+      "loss": 0.0551,
+      "step": 24420
+    },
+    {
+      "epoch": 21.92,
+      "learning_rate": 3.118890883702374e-05,
+      "loss": 0.0689,
+      "step": 24430
+    },
+    {
+      "epoch": 21.92,
+      "learning_rate": 3.1178934769599045e-05,
+      "loss": 0.0562,
+      "step": 24440
+    },
+    {
+      "epoch": 21.93,
+      "learning_rate": 3.116896070217435e-05,
+      "loss": 0.08,
+      "step": 24450
+    },
+    {
+      "epoch": 21.94,
+      "learning_rate": 3.115898663474965e-05,
+      "loss": 0.0643,
+      "step": 24460
+    },
+    {
+      "epoch": 21.95,
+      "learning_rate": 3.114901256732496e-05,
+      "loss": 0.0635,
+      "step": 24470
+    },
+    {
+      "epoch": 21.96,
+      "learning_rate": 3.113903849990026e-05,
+      "loss": 0.0645,
+      "step": 24480
+    },
+    {
+      "epoch": 21.97,
+      "learning_rate": 3.112906443247556e-05,
+      "loss": 0.0678,
+      "step": 24490
+    },
+    {
+      "epoch": 21.98,
+      "learning_rate": 3.111909036505087e-05,
+      "loss": 0.0674,
+      "step": 24500
+    },
+    {
+      "epoch": 21.99,
+      "learning_rate": 3.1109116297626175e-05,
+      "loss": 0.0456,
+      "step": 24510
+    },
+    {
+      "epoch": 22.0,
+      "learning_rate": 3.1099142230201475e-05,
+      "loss": 0.0753,
+      "step": 24520
+    },
+    {
+      "epoch": 22.0,
+      "eval_accuracy": {
+        "accuracy": 0.9798721686476789
+      },
+      "eval_f1": {
+        "f1": 0.9791539833000008
+      },
+      "eval_loss": 0.06731348484754562,
+      "eval_precision": {
+        "precision": 0.9788167338624194
+      },
+      "eval_recall": {
+        "recall": 0.9796378458227846
+      },
+      "eval_runtime": 185.3809,
+      "eval_samples_per_second": 96.213,
+      "eval_steps_per_second": 6.015,
+      "step": 24524
+    },
+    {
+      "epoch": 22.0,
+      "learning_rate": 3.108916816277678e-05,
+      "loss": 0.0466,
+      "step": 24530
+    },
+    {
+      "epoch": 22.01,
+      "learning_rate": 3.107919409535209e-05,
+      "loss": 0.0976,
+      "step": 24540
+    },
+    {
+      "epoch": 22.02,
+      "learning_rate": 3.106922002792739e-05,
+      "loss": 0.0623,
+      "step": 24550
+    },
+    {
+      "epoch": 22.03,
+      "learning_rate": 3.105924596050269e-05,
+      "loss": 0.0551,
+      "step": 24560
+    },
+    {
+      "epoch": 22.04,
+      "learning_rate": 3.1049271893078e-05,
+      "loss": 0.0452,
+      "step": 24570
+    },
+    {
+      "epoch": 22.05,
+      "learning_rate": 3.1039297825653305e-05,
+      "loss": 0.0769,
+      "step": 24580
+    },
+    {
+      "epoch": 22.06,
+      "learning_rate": 3.1029323758228604e-05,
+      "loss": 0.0795,
+      "step": 24590
+    },
+    {
+      "epoch": 22.07,
+      "learning_rate": 3.101934969080391e-05,
+      "loss": 0.0576,
+      "step": 24600
+    },
+    {
+      "epoch": 22.08,
+      "learning_rate": 3.100937562337922e-05,
+      "loss": 0.0731,
+      "step": 24610
+    },
+    {
+      "epoch": 22.09,
+      "learning_rate": 3.099940155595452e-05,
+      "loss": 0.061,
+      "step": 24620
+    },
+    {
+      "epoch": 22.09,
+      "learning_rate": 3.098942748852983e-05,
+      "loss": 0.0614,
+      "step": 24630
+    },
+    {
+      "epoch": 22.1,
+      "learning_rate": 3.097945342110513e-05,
+      "loss": 0.0753,
+      "step": 24640
+    },
+    {
+      "epoch": 22.11,
+      "learning_rate": 3.096947935368043e-05,
+      "loss": 0.0867,
+      "step": 24650
+    },
+    {
+      "epoch": 22.12,
+      "learning_rate": 3.0959505286255734e-05,
+      "loss": 0.0662,
+      "step": 24660
+    },
+    {
+      "epoch": 22.13,
+      "learning_rate": 3.094953121883104e-05,
+      "loss": 0.0712,
+      "step": 24670
+    },
+    {
+      "epoch": 22.14,
+      "learning_rate": 3.0939557151406347e-05,
+      "loss": 0.0621,
+      "step": 24680
+    },
+    {
+      "epoch": 22.15,
+      "learning_rate": 3.092958308398165e-05,
+      "loss": 0.0541,
+      "step": 24690
+    },
+    {
+      "epoch": 22.16,
+      "learning_rate": 3.091960901655695e-05,
+      "loss": 0.0596,
+      "step": 24700
+    },
+    {
+      "epoch": 22.17,
+      "learning_rate": 3.090963494913226e-05,
+      "loss": 0.1038,
+      "step": 24710
+    },
+    {
+      "epoch": 22.18,
+      "learning_rate": 3.089966088170756e-05,
+      "loss": 0.0734,
+      "step": 24720
+    },
+    {
+      "epoch": 22.18,
+      "learning_rate": 3.0889686814282864e-05,
+      "loss": 0.0662,
+      "step": 24730
+    },
+    {
+      "epoch": 22.19,
+      "learning_rate": 3.087971274685817e-05,
+      "loss": 0.0954,
+      "step": 24740
+    },
+    {
+      "epoch": 22.2,
+      "learning_rate": 3.0869738679433476e-05,
+      "loss": 0.0687,
+      "step": 24750
+    },
+    {
+      "epoch": 22.21,
+      "learning_rate": 3.0859764612008776e-05,
+      "loss": 0.0771,
+      "step": 24760
+    },
+    {
+      "epoch": 22.22,
+      "learning_rate": 3.084979054458408e-05,
+      "loss": 0.0837,
+      "step": 24770
+    },
+    {
+      "epoch": 22.23,
+      "learning_rate": 3.083981647715939e-05,
+      "loss": 0.1126,
+      "step": 24780
+    },
+    {
+      "epoch": 22.24,
+      "learning_rate": 3.0829842409734694e-05,
+      "loss": 0.0897,
+      "step": 24790
+    },
+    {
+      "epoch": 22.25,
+      "learning_rate": 3.0819868342309994e-05,
+      "loss": 0.0502,
+      "step": 24800
+    },
+    {
+      "epoch": 22.26,
+      "learning_rate": 3.08098942748853e-05,
+      "loss": 0.0396,
+      "step": 24810
+    },
+    {
+      "epoch": 22.27,
+      "learning_rate": 3.0799920207460606e-05,
+      "loss": 0.095,
+      "step": 24820
+    },
+    {
+      "epoch": 22.27,
+      "learning_rate": 3.0789946140035906e-05,
+      "loss": 0.0878,
+      "step": 24830
+    },
+    {
+      "epoch": 22.28,
+      "learning_rate": 3.077997207261121e-05,
+      "loss": 0.073,
+      "step": 24840
+    },
+    {
+      "epoch": 22.29,
+      "learning_rate": 3.076999800518652e-05,
+      "loss": 0.0529,
+      "step": 24850
+    },
+    {
+      "epoch": 22.3,
+      "learning_rate": 3.0760023937761824e-05,
+      "loss": 0.0728,
+      "step": 24860
+    },
+    {
+      "epoch": 22.31,
+      "learning_rate": 3.075004987033713e-05,
+      "loss": 0.0848,
+      "step": 24870
+    },
+    {
+      "epoch": 22.32,
+      "learning_rate": 3.074007580291243e-05,
+      "loss": 0.0644,
+      "step": 24880
+    },
+    {
+      "epoch": 22.33,
+      "learning_rate": 3.073010173548773e-05,
+      "loss": 0.0814,
+      "step": 24890
+    },
+    {
+      "epoch": 22.34,
+      "learning_rate": 3.0720127668063035e-05,
+      "loss": 0.0691,
+      "step": 24900
+    },
+    {
+      "epoch": 22.35,
+      "learning_rate": 3.071015360063834e-05,
+      "loss": 0.0931,
+      "step": 24910
+    },
+    {
+      "epoch": 22.35,
+      "learning_rate": 3.070017953321365e-05,
+      "loss": 0.0555,
+      "step": 24920
+    },
+    {
+      "epoch": 22.36,
+      "learning_rate": 3.0690205465788954e-05,
+      "loss": 0.0717,
+      "step": 24930
+    },
+    {
+      "epoch": 22.37,
+      "learning_rate": 3.068023139836425e-05,
+      "loss": 0.0646,
+      "step": 24940
+    },
+    {
+      "epoch": 22.38,
+      "learning_rate": 3.067025733093956e-05,
+      "loss": 0.0608,
+      "step": 24950
+    },
+    {
+      "epoch": 22.39,
+      "learning_rate": 3.066028326351486e-05,
+      "loss": 0.0842,
+      "step": 24960
+    },
+    {
+      "epoch": 22.4,
+      "learning_rate": 3.0650309196090165e-05,
+      "loss": 0.0589,
+      "step": 24970
+    },
+    {
+      "epoch": 22.41,
+      "learning_rate": 3.064033512866547e-05,
+      "loss": 0.0713,
+      "step": 24980
+    },
+    {
+      "epoch": 22.42,
+      "learning_rate": 3.063036106124078e-05,
+      "loss": 0.0509,
+      "step": 24990
+    },
+    {
+      "epoch": 22.43,
+      "learning_rate": 3.0620386993816084e-05,
+      "loss": 0.0917,
+      "step": 25000
+    },
+    {
+      "epoch": 22.44,
+      "learning_rate": 3.061041292639138e-05,
+      "loss": 0.0852,
+      "step": 25010
+    },
+    {
+      "epoch": 22.44,
+      "learning_rate": 3.060043885896669e-05,
+      "loss": 0.0675,
+      "step": 25020
+    },
+    {
+      "epoch": 22.45,
+      "learning_rate": 3.059046479154199e-05,
+      "loss": 0.0587,
+      "step": 25030
+    },
+    {
+      "epoch": 22.46,
+      "learning_rate": 3.0580490724117295e-05,
+      "loss": 0.0818,
+      "step": 25040
+    },
+    {
+      "epoch": 22.47,
+      "learning_rate": 3.05705166566926e-05,
+      "loss": 0.0599,
+      "step": 25050
+    },
+    {
+      "epoch": 22.48,
+      "learning_rate": 3.056054258926791e-05,
+      "loss": 0.0542,
+      "step": 25060
+    },
+    {
+      "epoch": 22.49,
+      "learning_rate": 3.055056852184321e-05,
+      "loss": 0.0498,
+      "step": 25070
+    },
+    {
+      "epoch": 22.5,
+      "learning_rate": 3.054059445441851e-05,
+      "loss": 0.0967,
+      "step": 25080
+    },
+    {
+      "epoch": 22.51,
+      "learning_rate": 3.053062038699382e-05,
+      "loss": 0.0751,
+      "step": 25090
+    },
+    {
+      "epoch": 22.52,
+      "learning_rate": 3.0520646319569125e-05,
+      "loss": 0.0777,
+      "step": 25100
+    },
+    {
+      "epoch": 22.53,
+      "learning_rate": 3.0510672252144425e-05,
+      "loss": 0.0405,
+      "step": 25110
+    },
+    {
+      "epoch": 22.53,
+      "learning_rate": 3.0500698184719728e-05,
+      "loss": 0.0757,
+      "step": 25120
+    },
+    {
+      "epoch": 22.54,
+      "learning_rate": 3.0490724117295034e-05,
+      "loss": 0.0713,
+      "step": 25130
+    },
+    {
+      "epoch": 22.55,
+      "learning_rate": 3.0480750049870337e-05,
+      "loss": 0.0565,
+      "step": 25140
+    },
+    {
+      "epoch": 22.56,
+      "learning_rate": 3.0470775982445643e-05,
+      "loss": 0.0762,
+      "step": 25150
+    },
+    {
+      "epoch": 22.57,
+      "learning_rate": 3.046080191502095e-05,
+      "loss": 0.0637,
+      "step": 25160
+    },
+    {
+      "epoch": 22.58,
+      "learning_rate": 3.0450827847596252e-05,
+      "loss": 0.0585,
+      "step": 25170
+    },
+    {
+      "epoch": 22.59,
+      "learning_rate": 3.0440853780171558e-05,
+      "loss": 0.0863,
+      "step": 25180
+    },
+    {
+      "epoch": 22.6,
+      "learning_rate": 3.0430879712746857e-05,
+      "loss": 0.0563,
+      "step": 25190
+    },
+    {
+      "epoch": 22.61,
+      "learning_rate": 3.0420905645322164e-05,
+      "loss": 0.0796,
+      "step": 25200
+    },
+    {
+      "epoch": 22.61,
+      "learning_rate": 3.0410931577897466e-05,
+      "loss": 0.065,
+      "step": 25210
+    },
+    {
+      "epoch": 22.62,
+      "learning_rate": 3.0400957510472772e-05,
+      "loss": 0.0783,
+      "step": 25220
+    },
+    {
+      "epoch": 22.63,
+      "learning_rate": 3.0390983443048075e-05,
+      "loss": 0.0772,
+      "step": 25230
+    },
+    {
+      "epoch": 22.64,
+      "learning_rate": 3.038100937562338e-05,
+      "loss": 0.051,
+      "step": 25240
+    },
+    {
+      "epoch": 22.65,
+      "learning_rate": 3.0371035308198688e-05,
+      "loss": 0.1123,
+      "step": 25250
+    },
+    {
+      "epoch": 22.66,
+      "learning_rate": 3.036106124077399e-05,
+      "loss": 0.0789,
+      "step": 25260
+    },
+    {
+      "epoch": 22.67,
+      "learning_rate": 3.035108717334929e-05,
+      "loss": 0.0598,
+      "step": 25270
+    },
+    {
+      "epoch": 22.68,
+      "learning_rate": 3.0341113105924596e-05,
+      "loss": 0.0931,
+      "step": 25280
+    },
+    {
+      "epoch": 22.69,
+      "learning_rate": 3.03311390384999e-05,
+      "loss": 0.0613,
+      "step": 25290
+    },
+    {
+      "epoch": 22.7,
+      "learning_rate": 3.0321164971075205e-05,
+      "loss": 0.0598,
+      "step": 25300
+    },
+    {
+      "epoch": 22.7,
+      "learning_rate": 3.031119090365051e-05,
+      "loss": 0.0758,
+      "step": 25310
+    },
+    {
+      "epoch": 22.71,
+      "learning_rate": 3.0301216836225814e-05,
+      "loss": 0.088,
+      "step": 25320
+    },
+    {
+      "epoch": 22.72,
+      "learning_rate": 3.029124276880112e-05,
+      "loss": 0.0582,
+      "step": 25330
+    },
+    {
+      "epoch": 22.73,
+      "learning_rate": 3.0281268701376426e-05,
+      "loss": 0.1118,
+      "step": 25340
+    },
+    {
+      "epoch": 22.74,
+      "learning_rate": 3.0271294633951726e-05,
+      "loss": 0.0698,
+      "step": 25350
+    },
+    {
+      "epoch": 22.75,
+      "learning_rate": 3.026132056652703e-05,
+      "loss": 0.0675,
+      "step": 25360
+    },
+    {
+      "epoch": 22.76,
+      "learning_rate": 3.0251346499102335e-05,
+      "loss": 0.0451,
+      "step": 25370
+    },
+    {
+      "epoch": 22.77,
+      "learning_rate": 3.0241372431677638e-05,
+      "loss": 0.0582,
+      "step": 25380
+    },
+    {
+      "epoch": 22.78,
+      "learning_rate": 3.0231398364252944e-05,
+      "loss": 0.0719,
+      "step": 25390
+    },
+    {
+      "epoch": 22.79,
+      "learning_rate": 3.022142429682825e-05,
+      "loss": 0.0577,
+      "step": 25400
+    },
+    {
+      "epoch": 22.79,
+      "learning_rate": 3.0211450229403553e-05,
+      "loss": 0.0773,
+      "step": 25410
+    },
+    {
+      "epoch": 22.8,
+      "learning_rate": 3.020147616197886e-05,
+      "loss": 0.0887,
+      "step": 25420
+    },
+    {
+      "epoch": 22.81,
+      "learning_rate": 3.019150209455416e-05,
+      "loss": 0.087,
+      "step": 25430
+    },
+    {
+      "epoch": 22.82,
+      "learning_rate": 3.0181528027129465e-05,
+      "loss": 0.05,
+      "step": 25440
+    },
+    {
+      "epoch": 22.83,
+      "learning_rate": 3.0171553959704767e-05,
+      "loss": 0.0885,
+      "step": 25450
+    },
+    {
+      "epoch": 22.84,
+      "learning_rate": 3.0161579892280074e-05,
+      "loss": 0.1027,
+      "step": 25460
+    },
+    {
+      "epoch": 22.85,
+      "learning_rate": 3.0151605824855376e-05,
+      "loss": 0.0619,
+      "step": 25470
+    },
+    {
+      "epoch": 22.86,
+      "learning_rate": 3.0141631757430683e-05,
+      "loss": 0.1015,
+      "step": 25480
+    },
+    {
+      "epoch": 22.87,
+      "learning_rate": 3.013165769000599e-05,
+      "loss": 0.0875,
+      "step": 25490
+    },
+    {
+      "epoch": 22.88,
+      "learning_rate": 3.012168362258129e-05,
+      "loss": 0.0871,
+      "step": 25500
+    },
+    {
+      "epoch": 22.88,
+      "learning_rate": 3.011170955515659e-05,
+      "loss": 0.0763,
+      "step": 25510
+    },
+    {
+      "epoch": 22.89,
+      "learning_rate": 3.0101735487731897e-05,
+      "loss": 0.079,
+      "step": 25520
+    },
+    {
+      "epoch": 22.9,
+      "learning_rate": 3.0091761420307203e-05,
+      "loss": 0.0716,
+      "step": 25530
+    },
+    {
+      "epoch": 22.91,
+      "learning_rate": 3.0081787352882506e-05,
+      "loss": 0.0657,
+      "step": 25540
+    },
+    {
+      "epoch": 22.92,
+      "learning_rate": 3.0071813285457812e-05,
+      "loss": 0.1025,
+      "step": 25550
+    },
+    {
+      "epoch": 22.93,
+      "learning_rate": 3.0061839218033115e-05,
+      "loss": 0.0712,
+      "step": 25560
+    },
+    {
+      "epoch": 22.94,
+      "learning_rate": 3.005186515060842e-05,
+      "loss": 0.0484,
+      "step": 25570
+    },
+    {
+      "epoch": 22.95,
+      "learning_rate": 3.0041891083183728e-05,
+      "loss": 0.0787,
+      "step": 25580
+    },
+    {
+      "epoch": 22.96,
+      "learning_rate": 3.0031917015759027e-05,
+      "loss": 0.0318,
+      "step": 25590
+    },
+    {
+      "epoch": 22.96,
+      "learning_rate": 3.002194294833433e-05,
+      "loss": 0.0645,
+      "step": 25600
+    },
+    {
+      "epoch": 22.97,
+      "learning_rate": 3.0011968880909636e-05,
+      "loss": 0.0713,
+      "step": 25610
+    },
+    {
+      "epoch": 22.98,
+      "learning_rate": 3.000199481348494e-05,
+      "loss": 0.0742,
+      "step": 25620
+    },
+    {
+      "epoch": 22.99,
+      "learning_rate": 2.9992020746060245e-05,
+      "loss": 0.0564,
+      "step": 25630
+    },
+    {
+      "epoch": 23.0,
+      "eval_accuracy": {
+        "accuracy": 0.9795357703520969
+      },
+      "eval_f1": {
+        "f1": 0.9785354369309488
+      },
+      "eval_loss": 0.06704173982143402,
+      "eval_precision": {
+        "precision": 0.9785832048371994
+      },
+      "eval_recall": {
+        "recall": 0.9785660480900121
+      },
+      "eval_runtime": 199.1143,
+      "eval_samples_per_second": 89.577,
+      "eval_steps_per_second": 5.6,
+      "step": 25639
+    },
+    {
+      "epoch": 23.0,
+      "learning_rate": 2.998204667863555e-05,
+      "loss": 0.0722,
+      "step": 25640
+    },
+    {
+      "epoch": 23.01,
+      "learning_rate": 2.9972072611210854e-05,
+      "loss": 0.0655,
+      "step": 25650
+    },
+    {
+      "epoch": 23.02,
+      "learning_rate": 2.996209854378616e-05,
+      "loss": 0.0494,
+      "step": 25660
+    },
+    {
+      "epoch": 23.03,
+      "learning_rate": 2.995212447636146e-05,
+      "loss": 0.0914,
+      "step": 25670
+    },
+    {
+      "epoch": 23.04,
+      "learning_rate": 2.9942150408936766e-05,
+      "loss": 0.0718,
+      "step": 25680
+    },
+    {
+      "epoch": 23.05,
+      "learning_rate": 2.993217634151207e-05,
+      "loss": 0.0732,
+      "step": 25690
+    },
+    {
+      "epoch": 23.05,
+      "learning_rate": 2.9922202274087375e-05,
+      "loss": 0.0974,
+      "step": 25700
+    },
+    {
+      "epoch": 23.06,
+      "learning_rate": 2.9912228206662678e-05,
+      "loss": 0.0748,
+      "step": 25710
+    },
+    {
+      "epoch": 23.07,
+      "learning_rate": 2.9902254139237984e-05,
+      "loss": 0.052,
+      "step": 25720
+    },
+    {
+      "epoch": 23.08,
+      "learning_rate": 2.989228007181329e-05,
+      "loss": 0.0769,
+      "step": 25730
+    },
+    {
+      "epoch": 23.09,
+      "learning_rate": 2.9882306004388593e-05,
+      "loss": 0.0757,
+      "step": 25740
+    },
+    {
+      "epoch": 23.1,
+      "learning_rate": 2.9872331936963892e-05,
+      "loss": 0.0594,
+      "step": 25750
+    },
+    {
+      "epoch": 23.11,
+      "learning_rate": 2.98623578695392e-05,
+      "loss": 0.0722,
+      "step": 25760
+    },
+    {
+      "epoch": 23.12,
+      "learning_rate": 2.9852383802114505e-05,
+      "loss": 0.0716,
+      "step": 25770
+    },
+    {
+      "epoch": 23.13,
+      "learning_rate": 2.9842409734689807e-05,
+      "loss": 0.046,
+      "step": 25780
+    },
+    {
+      "epoch": 23.14,
+      "learning_rate": 2.9832435667265114e-05,
+      "loss": 0.0537,
+      "step": 25790
+    },
+    {
+      "epoch": 23.14,
+      "learning_rate": 2.9822461599840416e-05,
+      "loss": 0.0767,
+      "step": 25800
+    },
+    {
+      "epoch": 23.15,
+      "learning_rate": 2.9812487532415723e-05,
+      "loss": 0.0597,
+      "step": 25810
+    },
+    {
+      "epoch": 23.16,
+      "learning_rate": 2.9802513464991022e-05,
+      "loss": 0.0502,
+      "step": 25820
+    },
+    {
+      "epoch": 23.17,
+      "learning_rate": 2.9792539397566328e-05,
+      "loss": 0.0821,
+      "step": 25830
+    },
+    {
+      "epoch": 23.18,
+      "learning_rate": 2.978256533014163e-05,
+      "loss": 0.074,
+      "step": 25840
+    },
+    {
+      "epoch": 23.19,
+      "learning_rate": 2.9772591262716937e-05,
+      "loss": 0.0561,
+      "step": 25850
+    },
+    {
+      "epoch": 23.2,
+      "learning_rate": 2.976261719529224e-05,
+      "loss": 0.0647,
+      "step": 25860
+    },
+    {
+      "epoch": 23.21,
+      "learning_rate": 2.9752643127867546e-05,
+      "loss": 0.071,
+      "step": 25870
+    },
+    {
+      "epoch": 23.22,
+      "learning_rate": 2.9742669060442852e-05,
+      "loss": 0.0814,
+      "step": 25880
+    },
+    {
+      "epoch": 23.22,
+      "learning_rate": 2.9732694993018155e-05,
+      "loss": 0.0598,
+      "step": 25890
+    },
+    {
+      "epoch": 23.23,
+      "learning_rate": 2.9722720925593455e-05,
+      "loss": 0.0518,
+      "step": 25900
+    },
+    {
+      "epoch": 23.24,
+      "learning_rate": 2.971274685816876e-05,
+      "loss": 0.0579,
+      "step": 25910
+    },
+    {
+      "epoch": 23.25,
+      "learning_rate": 2.9702772790744067e-05,
+      "loss": 0.0678,
+      "step": 25920
+    },
+    {
+      "epoch": 23.26,
+      "learning_rate": 2.969279872331937e-05,
+      "loss": 0.0845,
+      "step": 25930
+    },
+    {
+      "epoch": 23.27,
+      "learning_rate": 2.9682824655894676e-05,
+      "loss": 0.0839,
+      "step": 25940
+    },
+    {
+      "epoch": 23.28,
+      "learning_rate": 2.967285058846998e-05,
+      "loss": 0.0622,
+      "step": 25950
+    },
+    {
+      "epoch": 23.29,
+      "learning_rate": 2.9662876521045285e-05,
+      "loss": 0.0586,
+      "step": 25960
+    },
+    {
+      "epoch": 23.3,
+      "learning_rate": 2.965290245362059e-05,
+      "loss": 0.0653,
+      "step": 25970
+    },
+    {
+      "epoch": 23.31,
+      "learning_rate": 2.964292838619589e-05,
+      "loss": 0.053,
+      "step": 25980
+    },
+    {
+      "epoch": 23.31,
+      "learning_rate": 2.9632954318771193e-05,
+      "loss": 0.0517,
+      "step": 25990
+    },
+    {
+      "epoch": 23.32,
+      "learning_rate": 2.96229802513465e-05,
+      "loss": 0.0463,
+      "step": 26000
+    },
+    {
+      "epoch": 23.33,
+      "learning_rate": 2.9613006183921806e-05,
+      "loss": 0.0654,
+      "step": 26010
+    },
+    {
+      "epoch": 23.34,
+      "learning_rate": 2.960303211649711e-05,
+      "loss": 0.0673,
+      "step": 26020
+    },
+    {
+      "epoch": 23.35,
+      "learning_rate": 2.9593058049072415e-05,
+      "loss": 0.0827,
+      "step": 26030
+    },
+    {
+      "epoch": 23.36,
+      "learning_rate": 2.9583083981647718e-05,
+      "loss": 0.0659,
+      "step": 26040
+    },
+    {
+      "epoch": 23.37,
+      "learning_rate": 2.9573109914223024e-05,
+      "loss": 0.0562,
+      "step": 26050
+    },
+    {
+      "epoch": 23.38,
+      "learning_rate": 2.9563135846798323e-05,
+      "loss": 0.06,
+      "step": 26060
+    },
+    {
+      "epoch": 23.39,
+      "learning_rate": 2.955316177937363e-05,
+      "loss": 0.0706,
+      "step": 26070
+    },
+    {
+      "epoch": 23.4,
+      "learning_rate": 2.9543187711948932e-05,
+      "loss": 0.0651,
+      "step": 26080
+    },
+    {
+      "epoch": 23.4,
+      "learning_rate": 2.953321364452424e-05,
+      "loss": 0.058,
+      "step": 26090
+    },
+    {
+      "epoch": 23.41,
+      "learning_rate": 2.9523239577099545e-05,
+      "loss": 0.072,
+      "step": 26100
+    },
+    {
+      "epoch": 23.42,
+      "learning_rate": 2.9513265509674847e-05,
+      "loss": 0.0524,
+      "step": 26110
+    },
+    {
+      "epoch": 23.43,
+      "learning_rate": 2.9503291442250154e-05,
+      "loss": 0.1029,
+      "step": 26120
+    },
+    {
+      "epoch": 23.44,
+      "learning_rate": 2.9493317374825456e-05,
+      "loss": 0.0628,
+      "step": 26130
+    },
+    {
+      "epoch": 23.45,
+      "learning_rate": 2.9483343307400756e-05,
+      "loss": 0.0818,
+      "step": 26140
+    },
+    {
+      "epoch": 23.46,
+      "learning_rate": 2.9473369239976062e-05,
+      "loss": 0.0636,
+      "step": 26150
+    },
+    {
+      "epoch": 23.47,
+      "learning_rate": 2.9463395172551368e-05,
+      "loss": 0.0638,
+      "step": 26160
+    },
+    {
+      "epoch": 23.48,
+      "learning_rate": 2.945342110512667e-05,
+      "loss": 0.0548,
+      "step": 26170
+    },
+    {
+      "epoch": 23.49,
+      "learning_rate": 2.9443447037701977e-05,
+      "loss": 0.0726,
+      "step": 26180
+    },
+    {
+      "epoch": 23.49,
+      "learning_rate": 2.943347297027728e-05,
+      "loss": 0.0611,
+      "step": 26190
+    },
+    {
+      "epoch": 23.5,
+      "learning_rate": 2.9423498902852586e-05,
+      "loss": 0.049,
+      "step": 26200
+    },
+    {
+      "epoch": 23.51,
+      "learning_rate": 2.9413524835427892e-05,
+      "loss": 0.0527,
+      "step": 26210
+    },
+    {
+      "epoch": 23.52,
+      "learning_rate": 2.9403550768003192e-05,
+      "loss": 0.0604,
+      "step": 26220
+    },
+    {
+      "epoch": 23.53,
+      "learning_rate": 2.9393576700578495e-05,
+      "loss": 0.0843,
+      "step": 26230
+    },
+    {
+      "epoch": 23.54,
+      "learning_rate": 2.93836026331538e-05,
+      "loss": 0.0953,
+      "step": 26240
+    },
+    {
+      "epoch": 23.55,
+      "learning_rate": 2.9373628565729107e-05,
+      "loss": 0.1068,
+      "step": 26250
+    },
+    {
+      "epoch": 23.56,
+      "learning_rate": 2.936365449830441e-05,
+      "loss": 0.0717,
+      "step": 26260
+    },
+    {
+      "epoch": 23.57,
+      "learning_rate": 2.9353680430879716e-05,
+      "loss": 0.0794,
+      "step": 26270
+    },
+    {
+      "epoch": 23.57,
+      "learning_rate": 2.934370636345502e-05,
+      "loss": 0.0693,
+      "step": 26280
+    },
+    {
+      "epoch": 23.58,
+      "learning_rate": 2.9333732296030325e-05,
+      "loss": 0.0663,
+      "step": 26290
+    },
+    {
+      "epoch": 23.59,
+      "learning_rate": 2.9323758228605624e-05,
+      "loss": 0.0828,
+      "step": 26300
+    },
+    {
+      "epoch": 23.6,
+      "learning_rate": 2.931378416118093e-05,
+      "loss": 0.0577,
+      "step": 26310
+    },
+    {
+      "epoch": 23.61,
+      "learning_rate": 2.9303810093756233e-05,
+      "loss": 0.0458,
+      "step": 26320
+    },
+    {
+      "epoch": 23.62,
+      "learning_rate": 2.929383602633154e-05,
+      "loss": 0.0766,
+      "step": 26330
+    },
+    {
+      "epoch": 23.63,
+      "learning_rate": 2.9283861958906846e-05,
+      "loss": 0.0755,
+      "step": 26340
+    },
+    {
+      "epoch": 23.64,
+      "learning_rate": 2.927388789148215e-05,
+      "loss": 0.0577,
+      "step": 26350
+    },
+    {
+      "epoch": 23.65,
+      "learning_rate": 2.9263913824057455e-05,
+      "loss": 0.0924,
+      "step": 26360
+    },
+    {
+      "epoch": 23.66,
+      "learning_rate": 2.9253939756632758e-05,
+      "loss": 0.053,
+      "step": 26370
+    },
+    {
+      "epoch": 23.66,
+      "learning_rate": 2.9243965689208057e-05,
+      "loss": 0.1074,
+      "step": 26380
+    },
+    {
+      "epoch": 23.67,
+      "learning_rate": 2.9233991621783363e-05,
+      "loss": 0.0729,
+      "step": 26390
+    },
+    {
+      "epoch": 23.68,
+      "learning_rate": 2.922401755435867e-05,
+      "loss": 0.0627,
+      "step": 26400
+    },
+    {
+      "epoch": 23.69,
+      "learning_rate": 2.9214043486933972e-05,
+      "loss": 0.101,
+      "step": 26410
+    },
+    {
+      "epoch": 23.7,
+      "learning_rate": 2.920406941950928e-05,
+      "loss": 0.068,
+      "step": 26420
+    },
+    {
+      "epoch": 23.71,
+      "learning_rate": 2.919409535208458e-05,
+      "loss": 0.0688,
+      "step": 26430
+    },
+    {
+      "epoch": 23.72,
+      "learning_rate": 2.9184121284659887e-05,
+      "loss": 0.0714,
+      "step": 26440
+    },
+    {
+      "epoch": 23.73,
+      "learning_rate": 2.9174147217235194e-05,
+      "loss": 0.0701,
+      "step": 26450
+    },
+    {
+      "epoch": 23.74,
+      "learning_rate": 2.9164173149810493e-05,
+      "loss": 0.0707,
+      "step": 26460
+    },
+    {
+      "epoch": 23.75,
+      "learning_rate": 2.9154199082385796e-05,
+      "loss": 0.0612,
+      "step": 26470
+    },
+    {
+      "epoch": 23.75,
+      "learning_rate": 2.9144225014961102e-05,
+      "loss": 0.0778,
+      "step": 26480
+    },
+    {
+      "epoch": 23.76,
+      "learning_rate": 2.9134250947536408e-05,
+      "loss": 0.0776,
+      "step": 26490
+    },
+    {
+      "epoch": 23.77,
+      "learning_rate": 2.912427688011171e-05,
+      "loss": 0.0554,
+      "step": 26500
+    },
+    {
+      "epoch": 23.78,
+      "learning_rate": 2.9114302812687017e-05,
+      "loss": 0.0867,
+      "step": 26510
+    },
+    {
+      "epoch": 23.79,
+      "learning_rate": 2.910432874526232e-05,
+      "loss": 0.0701,
+      "step": 26520
+    },
+    {
+      "epoch": 23.8,
+      "learning_rate": 2.909435467783762e-05,
+      "loss": 0.0797,
+      "step": 26530
+    },
+    {
+      "epoch": 23.81,
+      "learning_rate": 2.9084380610412926e-05,
+      "loss": 0.0653,
+      "step": 26540
+    },
+    {
+      "epoch": 23.82,
+      "learning_rate": 2.9074406542988232e-05,
+      "loss": 0.0735,
+      "step": 26550
+    },
+    {
+      "epoch": 23.83,
+      "learning_rate": 2.9064432475563535e-05,
+      "loss": 0.0534,
+      "step": 26560
+    },
+    {
+      "epoch": 23.83,
+      "learning_rate": 2.905445840813884e-05,
+      "loss": 0.0719,
+      "step": 26570
+    },
+    {
+      "epoch": 23.84,
+      "learning_rate": 2.9044484340714147e-05,
+      "loss": 0.0976,
+      "step": 26580
+    },
+    {
+      "epoch": 23.85,
+      "learning_rate": 2.903451027328945e-05,
+      "loss": 0.0816,
+      "step": 26590
+    },
+    {
+      "epoch": 23.86,
+      "learning_rate": 2.9024536205864756e-05,
+      "loss": 0.0771,
+      "step": 26600
+    },
+    {
+      "epoch": 23.87,
+      "learning_rate": 2.9014562138440055e-05,
+      "loss": 0.0612,
+      "step": 26610
+    },
+    {
+      "epoch": 23.88,
+      "learning_rate": 2.9004588071015358e-05,
+      "loss": 0.0754,
+      "step": 26620
+    },
+    {
+      "epoch": 23.89,
+      "learning_rate": 2.8994614003590664e-05,
+      "loss": 0.0485,
+      "step": 26630
+    },
+    {
+      "epoch": 23.9,
+      "learning_rate": 2.898463993616597e-05,
+      "loss": 0.0553,
+      "step": 26640
+    },
+    {
+      "epoch": 23.91,
+      "learning_rate": 2.8974665868741273e-05,
+      "loss": 0.0821,
+      "step": 26650
+    },
+    {
+      "epoch": 23.92,
+      "learning_rate": 2.896469180131658e-05,
+      "loss": 0.081,
+      "step": 26660
+    },
+    {
+      "epoch": 23.92,
+      "learning_rate": 2.8954717733891886e-05,
+      "loss": 0.0589,
+      "step": 26670
+    },
+    {
+      "epoch": 23.93,
+      "learning_rate": 2.894474366646719e-05,
+      "loss": 0.0788,
+      "step": 26680
+    },
+    {
+      "epoch": 23.94,
+      "learning_rate": 2.8934769599042488e-05,
+      "loss": 0.0789,
+      "step": 26690
+    },
+    {
+      "epoch": 23.95,
+      "learning_rate": 2.8924795531617794e-05,
+      "loss": 0.0705,
+      "step": 26700
+    },
+    {
+      "epoch": 23.96,
+      "learning_rate": 2.8914821464193097e-05,
+      "loss": 0.057,
+      "step": 26710
+    },
+    {
+      "epoch": 23.97,
+      "learning_rate": 2.8904847396768403e-05,
+      "loss": 0.0482,
+      "step": 26720
+    },
+    {
+      "epoch": 23.98,
+      "learning_rate": 2.889487332934371e-05,
+      "loss": 0.058,
+      "step": 26730
+    },
+    {
+      "epoch": 23.99,
+      "learning_rate": 2.8884899261919012e-05,
+      "loss": 0.0572,
+      "step": 26740
+    },
+    {
+      "epoch": 24.0,
+      "learning_rate": 2.8874925194494318e-05,
+      "loss": 0.0721,
+      "step": 26750
+    },
+    {
+      "epoch": 24.0,
+      "eval_accuracy": {
+        "accuracy": 0.9800964341780668
+      },
+      "eval_f1": {
+        "f1": 0.9791442297494849
+      },
+      "eval_loss": 0.06854487210512161,
+      "eval_precision": {
+        "precision": 0.9800176266790614
+      },
+      "eval_recall": {
+        "recall": 0.9783165451081763
+      },
+      "eval_runtime": 189.516,
+      "eval_samples_per_second": 94.113,
+      "eval_steps_per_second": 5.883,
+      "step": 26754
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 55700,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 50,
+  "save_steps": 500,
+  "total_flos": 1.3268106325269283e+20,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}