diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,7077 @@
+{
+  "best_metric": 0.993526599442064,
+  "best_model_checkpoint": "output_dev/v34-deberta_large_v3-f1-warmup-greater-patience-seqlen-240-6epoch/checkpoint-6600",
+  "epoch": 3.1755542240862793,
+  "eval_steps": 200,
+  "global_step": 10600,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.990019960079841e-07,
+      "loss": 1.9756,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.980039920159682e-07,
+      "loss": 1.8411,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.4970059880239521e-06,
+      "loss": 1.5232,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9960079840319363e-06,
+      "loss": 1.0963,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.4950099800399203e-06,
+      "loss": 0.7066,
+      "step": 50
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.9940119760479042e-06,
+      "loss": 0.5216,
+      "step": 60
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.493013972055888e-06,
+      "loss": 0.4144,
+      "step": 70
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.992015968063873e-06,
+      "loss": 0.3314,
+      "step": 80
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.4910179640718566e-06,
+      "loss": 0.2475,
+      "step": 90
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9900199600798405e-06,
+      "loss": 0.2203,
+      "step": 100
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.4890219560878245e-06,
+      "loss": 0.2609,
+      "step": 110
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.9880239520958085e-06,
+      "loss": 0.2257,
+      "step": 120
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.4870259481037925e-06,
+      "loss": 0.1866,
+      "step": 130
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.986027944111776e-06,
+      "loss": 0.144,
+      "step": 140
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 7.48502994011976e-06,
+      "loss": 0.1259,
+      "step": 150
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 7.984031936127745e-06,
+      "loss": 0.1448,
+      "step": 160
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 8.483033932135728e-06,
+      "loss": 0.1018,
+      "step": 170
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 8.982035928143713e-06,
+      "loss": 0.1446,
+      "step": 180
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.481037924151696e-06,
+      "loss": 0.1391,
+      "step": 190
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 9.980039920159681e-06,
+      "loss": 0.0954,
+      "step": 200
+    },
+    {
+      "epoch": 0.06,
+      "eval_f1": 0.9439515816965873,
+      "eval_f1_0": 0.9814644883052654,
+      "eval_f1_1": 0.9883150101831029,
+      "eval_f1_2": 0.8092399020555978,
+      "eval_f1_3": 0.9967869262423833,
+      "eval_loss": 0.07926841825246811,
+      "eval_runtime": 52.7827,
+      "eval_samples_per_second": 224.846,
+      "eval_steps_per_second": 0.89,
+      "step": 200
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.0479041916167664e-05,
+      "loss": 0.0892,
+      "step": 210
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.0978043912175649e-05,
+      "loss": 0.0809,
+      "step": 220
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.1477045908183632e-05,
+      "loss": 0.0819,
+      "step": 230
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.1976047904191617e-05,
+      "loss": 0.0607,
+      "step": 240
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.24750499001996e-05,
+      "loss": 0.0641,
+      "step": 250
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.2974051896207585e-05,
+      "loss": 0.0579,
+      "step": 260
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.347305389221557e-05,
+      "loss": 0.0686,
+      "step": 270
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3972055888223553e-05,
+      "loss": 0.0353,
+      "step": 280
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.4471057884231538e-05,
+      "loss": 0.0894,
+      "step": 290
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.497005988023952e-05,
+      "loss": 0.0411,
+      "step": 300
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.5469061876247507e-05,
+      "loss": 0.0391,
+      "step": 310
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.596806387225549e-05,
+      "loss": 0.0548,
+      "step": 320
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.6467065868263474e-05,
+      "loss": 0.0768,
+      "step": 330
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.6966067864271457e-05,
+      "loss": 0.0373,
+      "step": 340
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.7465069860279443e-05,
+      "loss": 0.0519,
+      "step": 350
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.7964071856287426e-05,
+      "loss": 0.0399,
+      "step": 360
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.846307385229541e-05,
+      "loss": 0.0972,
+      "step": 370
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.8962075848303393e-05,
+      "loss": 0.0354,
+      "step": 380
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.946107784431138e-05,
+      "loss": 0.0827,
+      "step": 390
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9960079840319362e-05,
+      "loss": 0.0854,
+      "step": 400
+    },
+    {
+      "epoch": 0.12,
+      "eval_f1": 0.9588132852653405,
+      "eval_f1_0": 0.9854538374890487,
+      "eval_f1_1": 0.9896395970160494,
+      "eval_f1_2": 0.8628368371448905,
+      "eval_f1_3": 0.9973228694113733,
+      "eval_loss": 0.08576956391334534,
+      "eval_runtime": 53.1054,
+      "eval_samples_per_second": 223.48,
+      "eval_steps_per_second": 0.885,
+      "step": 400
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.0459081836327345e-05,
+      "loss": 0.0434,
+      "step": 410
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.095808383233533e-05,
+      "loss": 0.0715,
+      "step": 420
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.1457085828343315e-05,
+      "loss": 0.0385,
+      "step": 430
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.1956087824351298e-05,
+      "loss": 0.0474,
+      "step": 440
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.245508982035928e-05,
+      "loss": 0.0997,
+      "step": 450
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.2954091816367264e-05,
+      "loss": 0.0355,
+      "step": 460
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.345309381237525e-05,
+      "loss": 0.0515,
+      "step": 470
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.3952095808383234e-05,
+      "loss": 0.041,
+      "step": 480
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.4451097804391217e-05,
+      "loss": 0.0433,
+      "step": 490
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.49500998003992e-05,
+      "loss": 0.0308,
+      "step": 500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.5449101796407187e-05,
+      "loss": 0.0449,
+      "step": 510
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.594810379241517e-05,
+      "loss": 0.0524,
+      "step": 520
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6447105788423153e-05,
+      "loss": 0.0551,
+      "step": 530
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.694610778443114e-05,
+      "loss": 0.0535,
+      "step": 540
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.7445109780439123e-05,
+      "loss": 0.0539,
+      "step": 550
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.7944111776447106e-05,
+      "loss": 0.1137,
+      "step": 560
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8443113772455092e-05,
+      "loss": 0.0549,
+      "step": 570
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8942115768463075e-05,
+      "loss": 0.0349,
+      "step": 580
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.944111776447106e-05,
+      "loss": 0.0721,
+      "step": 590
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.994011976047904e-05,
+      "loss": 0.0206,
+      "step": 600
+    },
+    {
+      "epoch": 0.18,
+      "eval_f1": 0.9834921014335102,
+      "eval_f1_0": 0.9949107875077299,
+      "eval_f1_1": 0.9961045718004788,
+      "eval_f1_2": 0.9451365879813575,
+      "eval_f1_3": 0.9978164584444748,
+      "eval_loss": 0.028582798317074776,
+      "eval_runtime": 53.2231,
+      "eval_samples_per_second": 222.986,
+      "eval_steps_per_second": 0.883,
+      "step": 600
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 3.0439121756487028e-05,
+      "loss": 0.0223,
+      "step": 610
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 3.0938123752495015e-05,
+      "loss": 0.1399,
+      "step": 620
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 3.143712574850299e-05,
+      "loss": 0.0508,
+      "step": 630
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 3.193612774451098e-05,
+      "loss": 0.0129,
+      "step": 640
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 3.2435129740518964e-05,
+      "loss": 0.0671,
+      "step": 650
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.293413173652695e-05,
+      "loss": 0.0246,
+      "step": 660
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.343313373253493e-05,
+      "loss": 0.0491,
+      "step": 670
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.393213572854291e-05,
+      "loss": 0.0762,
+      "step": 680
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 3.4431137724550896e-05,
+      "loss": 0.0414,
+      "step": 690
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 3.4930139720558886e-05,
+      "loss": 0.0695,
+      "step": 700
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 3.542914171656686e-05,
+      "loss": 0.0396,
+      "step": 710
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 3.592814371257485e-05,
+      "loss": 0.0526,
+      "step": 720
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 3.6427145708582836e-05,
+      "loss": 0.0244,
+      "step": 730
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 3.692614770459082e-05,
+      "loss": 0.04,
+      "step": 740
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 3.74251497005988e-05,
+      "loss": 0.0425,
+      "step": 750
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.7924151696606785e-05,
+      "loss": 0.0213,
+      "step": 760
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.8423153692614775e-05,
+      "loss": 0.0424,
+      "step": 770
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.892215568862276e-05,
+      "loss": 0.0457,
+      "step": 780
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.942115768463074e-05,
+      "loss": 0.0576,
+      "step": 790
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.9920159680638724e-05,
+      "loss": 0.1332,
+      "step": 800
+    },
+    {
+      "epoch": 0.24,
+      "eval_f1": 0.9738886081580435,
+      "eval_f1_0": 0.9914198678301208,
+      "eval_f1_1": 0.9887483743062586,
+      "eval_f1_2": 0.9229570008505884,
+      "eval_f1_3": 0.9924291896452061,
+      "eval_loss": 0.04060590639710426,
+      "eval_runtime": 53.3557,
+      "eval_samples_per_second": 222.432,
+      "eval_steps_per_second": 0.881,
+      "step": 800
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.041916167664671e-05,
+      "loss": 0.0437,
+      "step": 810
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.091816367265469e-05,
+      "loss": 0.0643,
+      "step": 820
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.141716566866268e-05,
+      "loss": 0.0208,
+      "step": 830
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.191616766467066e-05,
+      "loss": 0.0252,
+      "step": 840
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.241516966067865e-05,
+      "loss": 0.0494,
+      "step": 850
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.291417165668663e-05,
+      "loss": 0.0167,
+      "step": 860
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.341317365269461e-05,
+      "loss": 0.0499,
+      "step": 870
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.3912175648702596e-05,
+      "loss": 0.0649,
+      "step": 880
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.4411177644710586e-05,
+      "loss": 0.0572,
+      "step": 890
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.491017964071856e-05,
+      "loss": 0.0508,
+      "step": 900
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.540918163672655e-05,
+      "loss": 0.1019,
+      "step": 910
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.590818363273453e-05,
+      "loss": 0.0529,
+      "step": 920
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.640718562874252e-05,
+      "loss": 0.0506,
+      "step": 930
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.69061876247505e-05,
+      "loss": 0.0767,
+      "step": 940
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7405189620758485e-05,
+      "loss": 0.0351,
+      "step": 950
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.790419161676647e-05,
+      "loss": 0.038,
+      "step": 960
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.840319361277446e-05,
+      "loss": 0.0589,
+      "step": 970
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.8902195608782434e-05,
+      "loss": 0.0338,
+      "step": 980
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9401197604790424e-05,
+      "loss": 0.0593,
+      "step": 990
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.99001996007984e-05,
+      "loss": 0.0263,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3,
+      "eval_f1": 0.9540581603427759,
+      "eval_f1_0": 0.9875224442592623,
+      "eval_f1_1": 0.996863689344212,
+      "eval_f1_2": 0.8345400374073041,
+      "eval_f1_3": 0.9973064703603255,
+      "eval_loss": 0.06365196406841278,
+      "eval_runtime": 53.2716,
+      "eval_samples_per_second": 222.783,
+      "eval_steps_per_second": 0.882,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.997897613791654e-05,
+      "loss": 0.0399,
+      "step": 1010
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.995269631031221e-05,
+      "loss": 0.0461,
+      "step": 1020
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9926416482707875e-05,
+      "loss": 0.0289,
+      "step": 1030
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.990013665510355e-05,
+      "loss": 0.033,
+      "step": 1040
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.987385682749922e-05,
+      "loss": 0.0549,
+      "step": 1050
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9847576999894884e-05,
+      "loss": 0.0373,
+      "step": 1060
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9821297172290556e-05,
+      "loss": 0.0359,
+      "step": 1070
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.979501734468622e-05,
+      "loss": 0.1199,
+      "step": 1080
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.976873751708189e-05,
+      "loss": 0.0546,
+      "step": 1090
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9742457689477564e-05,
+      "loss": 0.0299,
+      "step": 1100
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.971617786187323e-05,
+      "loss": 0.0481,
+      "step": 1110
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.96898980342689e-05,
+      "loss": 0.0205,
+      "step": 1120
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.9663618206664566e-05,
+      "loss": 0.0522,
+      "step": 1130
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.963733837906023e-05,
+      "loss": 0.0449,
+      "step": 1140
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.96110585514559e-05,
+      "loss": 0.0337,
+      "step": 1150
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9584778723851575e-05,
+      "loss": 0.0085,
+      "step": 1160
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.955849889624724e-05,
+      "loss": 0.1178,
+      "step": 1170
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.953221906864291e-05,
+      "loss": 0.0294,
+      "step": 1180
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.9505939241038577e-05,
+      "loss": 0.0229,
+      "step": 1190
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.947965941343425e-05,
+      "loss": 0.0268,
+      "step": 1200
+    },
+    {
+      "epoch": 0.36,
+      "eval_f1": 0.9767413552749206,
+      "eval_f1_0": 0.9923689597585356,
+      "eval_f1_1": 0.9898508596941199,
+      "eval_f1_2": 0.9286801598218379,
+      "eval_f1_3": 0.9960654418251892,
+      "eval_loss": 0.04390137270092964,
+      "eval_runtime": 53.2749,
+      "eval_samples_per_second": 222.769,
+      "eval_steps_per_second": 0.882,
+      "step": 1200
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.945337958582992e-05,
+      "loss": 0.0236,
+      "step": 1210
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9427099758225585e-05,
+      "loss": 0.0315,
+      "step": 1220
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.940081993062126e-05,
+      "loss": 0.024,
+      "step": 1230
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.937454010301692e-05,
+      "loss": 0.0675,
+      "step": 1240
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.9348260275412594e-05,
+      "loss": 0.0241,
+      "step": 1250
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.9321980447808266e-05,
+      "loss": 0.0489,
+      "step": 1260
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.929570062020393e-05,
+      "loss": 0.0322,
+      "step": 1270
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.92694207925996e-05,
+      "loss": 0.03,
+      "step": 1280
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.9243140964995274e-05,
+      "loss": 0.0488,
+      "step": 1290
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.921686113739094e-05,
+      "loss": 0.0417,
+      "step": 1300
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.919058130978661e-05,
+      "loss": 0.0283,
+      "step": 1310
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9164301482182276e-05,
+      "loss": 0.0441,
+      "step": 1320
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.913802165457795e-05,
+      "loss": 0.0279,
+      "step": 1330
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.911174182697362e-05,
+      "loss": 0.1011,
+      "step": 1340
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9085461999369285e-05,
+      "loss": 0.0374,
+      "step": 1350
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.9059182171764956e-05,
+      "loss": 0.0632,
+      "step": 1360
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.903290234416063e-05,
+      "loss": 0.0235,
+      "step": 1370
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.900662251655629e-05,
+      "loss": 0.0326,
+      "step": 1380
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.8980342688951965e-05,
+      "loss": 0.0333,
+      "step": 1390
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.895406286134763e-05,
+      "loss": 0.0285,
+      "step": 1400
+    },
+    {
+      "epoch": 0.42,
+      "eval_f1": 0.9798388589684135,
+      "eval_f1_0": 0.9935598148642307,
+      "eval_f1_1": 0.9969184361733485,
+      "eval_f1_2": 0.9313853930948695,
+      "eval_f1_3": 0.997491791741205,
+      "eval_loss": 0.03308385610580444,
+      "eval_runtime": 53.2818,
+      "eval_samples_per_second": 222.74,
+      "eval_steps_per_second": 0.882,
+      "step": 1400
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.89277830337433e-05,
+      "loss": 0.078,
+      "step": 1410
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8901503206138974e-05,
+      "loss": 0.0296,
+      "step": 1420
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.887522337853464e-05,
+      "loss": 0.0189,
+      "step": 1430
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.884894355093031e-05,
+      "loss": 0.0778,
+      "step": 1440
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8822663723325975e-05,
+      "loss": 0.0461,
+      "step": 1450
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.879638389572165e-05,
+      "loss": 0.0226,
+      "step": 1460
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.877010406811732e-05,
+      "loss": 0.0578,
+      "step": 1470
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8743824240512984e-05,
+      "loss": 0.0235,
+      "step": 1480
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8717544412908656e-05,
+      "loss": 0.0191,
+      "step": 1490
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.869126458530433e-05,
+      "loss": 0.0691,
+      "step": 1500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.866498475769999e-05,
+      "loss": 0.0303,
+      "step": 1510
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8638704930095664e-05,
+      "loss": 0.0321,
+      "step": 1520
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.861242510249133e-05,
+      "loss": 0.03,
+      "step": 1530
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8586145274887e-05,
+      "loss": 0.0415,
+      "step": 1540
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.855986544728267e-05,
+      "loss": 0.0396,
+      "step": 1550
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.853358561967833e-05,
+      "loss": 0.0186,
+      "step": 1560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8507305792074e-05,
+      "loss": 0.0275,
+      "step": 1570
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.8481025964469675e-05,
+      "loss": 0.0633,
+      "step": 1580
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.845474613686534e-05,
+      "loss": 0.0265,
+      "step": 1590
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.842846630926101e-05,
+      "loss": 0.048,
+      "step": 1600
+    },
+    {
+      "epoch": 0.48,
+      "eval_f1": 0.96749762324556,
+      "eval_f1_0": 0.9887758866484124,
+      "eval_f1_1": 0.9954477523516124,
+      "eval_f1_2": 0.8874648365802215,
+      "eval_f1_3": 0.9983020174019935,
+      "eval_loss": 0.05787326395511627,
+      "eval_runtime": 53.2266,
+      "eval_samples_per_second": 222.971,
+      "eval_steps_per_second": 0.883,
+      "step": 1600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8402186481656684e-05,
+      "loss": 0.0518,
+      "step": 1610
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.837590665405235e-05,
+      "loss": 0.0659,
+      "step": 1620
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.834962682644802e-05,
+      "loss": 0.032,
+      "step": 1630
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8323346998843685e-05,
+      "loss": 0.0248,
+      "step": 1640
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.829706717123936e-05,
+      "loss": 0.0347,
+      "step": 1650
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.827078734363503e-05,
+      "loss": 0.0086,
+      "step": 1660
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8244507516030694e-05,
+      "loss": 0.0196,
+      "step": 1670
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8218227688426366e-05,
+      "loss": 0.0692,
+      "step": 1680
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.819194786082204e-05,
+      "loss": 0.0225,
+      "step": 1690
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.81656680332177e-05,
+      "loss": 0.0059,
+      "step": 1700
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8139388205613374e-05,
+      "loss": 0.0389,
+      "step": 1710
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.811310837800904e-05,
+      "loss": 0.0111,
+      "step": 1720
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.808682855040471e-05,
+      "loss": 0.0143,
+      "step": 1730
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.806054872280038e-05,
+      "loss": 0.0269,
+      "step": 1740
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.803426889519605e-05,
+      "loss": 0.014,
+      "step": 1750
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.800798906759172e-05,
+      "loss": 0.0511,
+      "step": 1760
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.7981709239987385e-05,
+      "loss": 0.0541,
+      "step": 1770
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.795542941238306e-05,
+      "loss": 0.0225,
+      "step": 1780
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.792914958477873e-05,
+      "loss": 0.0277,
+      "step": 1790
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.7902869757174393e-05,
+      "loss": 0.0405,
+      "step": 1800
+    },
+    {
+      "epoch": 0.54,
+      "eval_f1": 0.9814626692863833,
+      "eval_f1_0": 0.9941325744122351,
+      "eval_f1_1": 0.9968333199151974,
+      "eval_f1_2": 0.936212977156517,
+      "eval_f1_3": 0.9986718056615835,
+      "eval_loss": 0.03122573159635067,
+      "eval_runtime": 53.2789,
+      "eval_samples_per_second": 222.752,
+      "eval_steps_per_second": 0.882,
+      "step": 1800
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.7876589929570065e-05,
+      "loss": 0.0296,
+      "step": 1810
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.785031010196574e-05,
+      "loss": 0.0165,
+      "step": 1820
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.78240302743614e-05,
+      "loss": 0.0221,
+      "step": 1830
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.7797750446757074e-05,
+      "loss": 0.0557,
+      "step": 1840
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.777147061915274e-05,
+      "loss": 0.063,
+      "step": 1850
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.774519079154841e-05,
+      "loss": 0.0425,
+      "step": 1860
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.771891096394408e-05,
+      "loss": 0.0653,
+      "step": 1870
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.769263113633975e-05,
+      "loss": 0.0647,
+      "step": 1880
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.766635130873542e-05,
+      "loss": 0.0344,
+      "step": 1890
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.764007148113109e-05,
+      "loss": 0.0473,
+      "step": 1900
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.7613791653526756e-05,
+      "loss": 0.0264,
+      "step": 1910
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.758751182592243e-05,
+      "loss": 0.0387,
+      "step": 1920
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.756123199831809e-05,
+      "loss": 0.0176,
+      "step": 1930
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.7534952170713765e-05,
+      "loss": 0.0394,
+      "step": 1940
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.7508672343109436e-05,
+      "loss": 0.062,
+      "step": 1950
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.74823925155051e-05,
+      "loss": 0.025,
+      "step": 1960
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.745611268790077e-05,
+      "loss": 0.0155,
+      "step": 1970
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.742983286029644e-05,
+      "loss": 0.0777,
+      "step": 1980
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.74035530326921e-05,
+      "loss": 0.0351,
+      "step": 1990
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.7377273205087775e-05,
+      "loss": 0.0303,
+      "step": 2000
+    },
+    {
+      "epoch": 0.6,
+      "eval_f1": 0.9755338337989246,
+      "eval_f1_0": 0.991693038431387,
+      "eval_f1_1": 0.9877976376772832,
+      "eval_f1_2": 0.9242240359623443,
+      "eval_f1_3": 0.9984206231246839,
+      "eval_loss": 0.04797644168138504,
+      "eval_runtime": 53.2786,
+      "eval_samples_per_second": 222.754,
+      "eval_steps_per_second": 0.882,
+      "step": 2000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.735099337748345e-05,
+      "loss": 0.0407,
+      "step": 2010
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.732471354987911e-05,
+      "loss": 0.0417,
+      "step": 2020
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.7298433722274784e-05,
+      "loss": 0.0374,
+      "step": 2030
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.727215389467045e-05,
+      "loss": 0.023,
+      "step": 2040
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.724587406706612e-05,
+      "loss": 0.0271,
+      "step": 2050
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.721959423946179e-05,
+      "loss": 0.0305,
+      "step": 2060
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.719331441185746e-05,
+      "loss": 0.0368,
+      "step": 2070
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.716703458425313e-05,
+      "loss": 0.0251,
+      "step": 2080
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.7140754756648794e-05,
+      "loss": 0.077,
+      "step": 2090
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.7114474929044466e-05,
+      "loss": 0.0095,
+      "step": 2100
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.708819510144014e-05,
+      "loss": 0.0311,
+      "step": 2110
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.70619152738358e-05,
+      "loss": 0.053,
+      "step": 2120
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.7035635446231475e-05,
+      "loss": 0.0396,
+      "step": 2130
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.7009355618627146e-05,
+      "loss": 0.0376,
+      "step": 2140
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.698307579102281e-05,
+      "loss": 0.0305,
+      "step": 2150
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.695679596341848e-05,
+      "loss": 0.0246,
+      "step": 2160
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.693051613581415e-05,
+      "loss": 0.0169,
+      "step": 2170
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.690423630820982e-05,
+      "loss": 0.037,
+      "step": 2180
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.687795648060549e-05,
+      "loss": 0.0289,
+      "step": 2190
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.685167665300116e-05,
+      "loss": 0.0389,
+      "step": 2200
+    },
+    {
+      "epoch": 0.66,
+      "eval_f1": 0.9832259546184603,
+      "eval_f1_0": 0.9946489155319365,
+      "eval_f1_1": 0.9966030472889557,
+      "eval_f1_2": 0.942989870485977,
+      "eval_f1_3": 0.9986619851669725,
+      "eval_loss": 0.027967283502221107,
+      "eval_runtime": 53.2558,
+      "eval_samples_per_second": 222.849,
+      "eval_steps_per_second": 0.883,
+      "step": 2200
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.682539682539683e-05,
+      "loss": 0.0477,
+      "step": 2210
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.67991169977925e-05,
+      "loss": 0.029,
+      "step": 2220
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.6772837170188165e-05,
+      "loss": 0.0405,
+      "step": 2230
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.674655734258384e-05,
+      "loss": 0.0305,
+      "step": 2240
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.67202775149795e-05,
+      "loss": 0.0323,
+      "step": 2250
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.6693997687375174e-05,
+      "loss": 0.0155,
+      "step": 2260
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.6667717859770846e-05,
+      "loss": 0.0205,
+      "step": 2270
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.664143803216651e-05,
+      "loss": 0.0333,
+      "step": 2280
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.661515820456218e-05,
+      "loss": 0.0231,
+      "step": 2290
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.658887837695785e-05,
+      "loss": 0.0875,
+      "step": 2300
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.656259854935352e-05,
+      "loss": 0.0398,
+      "step": 2310
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.653631872174919e-05,
+      "loss": 0.0253,
+      "step": 2320
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.6510038894144856e-05,
+      "loss": 0.039,
+      "step": 2330
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.648375906654053e-05,
+      "loss": 0.0173,
+      "step": 2340
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.64574792389362e-05,
+      "loss": 0.0143,
+      "step": 2350
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.6431199411331865e-05,
+      "loss": 0.0337,
+      "step": 2360
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.640491958372754e-05,
+      "loss": 0.0312,
+      "step": 2370
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.63786397561232e-05,
+      "loss": 0.0258,
+      "step": 2380
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.6352359928518873e-05,
+      "loss": 0.0154,
+      "step": 2390
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.6326080100914545e-05,
+      "loss": 0.0408,
+      "step": 2400
+    },
+    {
+      "epoch": 0.72,
+      "eval_f1": 0.9872307516679847,
+      "eval_f1_0": 0.9959209993448118,
+      "eval_f1_1": 0.9971936532504775,
+      "eval_f1_2": 0.9570808398310161,
+      "eval_f1_3": 0.9987275142456332,
+      "eval_loss": 0.0252233799546957,
+      "eval_runtime": 53.2552,
+      "eval_samples_per_second": 222.851,
+      "eval_steps_per_second": 0.883,
+      "step": 2400
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.6299800273310204e-05,
+      "loss": 0.0187,
+      "step": 2410
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.6273520445705875e-05,
+      "loss": 0.0048,
+      "step": 2420
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.624724061810155e-05,
+      "loss": 0.0156,
+      "step": 2430
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.622096079049721e-05,
+      "loss": 0.008,
+      "step": 2440
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.6194680962892884e-05,
+      "loss": 0.0582,
+      "step": 2450
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.6168401135288556e-05,
+      "loss": 0.0024,
+      "step": 2460
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.614212130768422e-05,
+      "loss": 0.0291,
+      "step": 2470
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.611584148007989e-05,
+      "loss": 0.0394,
+      "step": 2480
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.608956165247556e-05,
+      "loss": 0.0184,
+      "step": 2490
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.606328182487123e-05,
+      "loss": 0.0401,
+      "step": 2500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.60370019972669e-05,
+      "loss": 0.0456,
+      "step": 2510
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.6010722169662566e-05,
+      "loss": 0.0229,
+      "step": 2520
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.598444234205824e-05,
+      "loss": 0.0277,
+      "step": 2530
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.59581625144539e-05,
+      "loss": 0.025,
+      "step": 2540
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.5931882686849575e-05,
+      "loss": 0.0149,
+      "step": 2550
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.5905602859245247e-05,
+      "loss": 0.017,
+      "step": 2560
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.587932303164091e-05,
+      "loss": 0.0231,
+      "step": 2570
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.585304320403658e-05,
+      "loss": 0.0129,
+      "step": 2580
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.5826763376432255e-05,
+      "loss": 0.0147,
+      "step": 2590
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.580048354882792e-05,
+      "loss": 0.0196,
+      "step": 2600
+    },
+    {
+      "epoch": 0.78,
+      "eval_f1": 0.9887726640579566,
+      "eval_f1_0": 0.9963991401714185,
+      "eval_f1_1": 0.9967365610435348,
+      "eval_f1_2": 0.9631465838244995,
+      "eval_f1_3": 0.9988083711923735,
+      "eval_loss": 0.02168872021138668,
+      "eval_runtime": 53.355,
+      "eval_samples_per_second": 222.435,
+      "eval_steps_per_second": 0.881,
+      "step": 2600
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.577420372122359e-05,
+      "loss": 0.0072,
+      "step": 2610
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.574792389361926e-05,
+      "loss": 0.03,
+      "step": 2620
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.572164406601493e-05,
+      "loss": 0.0481,
+      "step": 2630
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.56953642384106e-05,
+      "loss": 0.08,
+      "step": 2640
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.5669084410806266e-05,
+      "loss": 0.0629,
+      "step": 2650
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.564280458320194e-05,
+      "loss": 0.038,
+      "step": 2660
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.561652475559761e-05,
+      "loss": 0.0187,
+      "step": 2670
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.5590244927993274e-05,
+      "loss": 0.0428,
+      "step": 2680
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.5563965100388946e-05,
+      "loss": 0.0157,
+      "step": 2690
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.553768527278461e-05,
+      "loss": 0.0149,
+      "step": 2700
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.551140544518028e-05,
+      "loss": 0.0224,
+      "step": 2710
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.5485125617575955e-05,
+      "loss": 0.0391,
+      "step": 2720
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.545884578997162e-05,
+      "loss": 0.0365,
+      "step": 2730
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.543256596236729e-05,
+      "loss": 0.0371,
+      "step": 2740
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.5406286134762956e-05,
+      "loss": 0.0165,
+      "step": 2750
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.538000630715863e-05,
+      "loss": 0.0354,
+      "step": 2760
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.53537264795543e-05,
+      "loss": 0.0225,
+      "step": 2770
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.5327446651949965e-05,
+      "loss": 0.0552,
+      "step": 2780
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.530116682434564e-05,
+      "loss": 0.0293,
+      "step": 2790
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.527488699674131e-05,
+      "loss": 0.039,
+      "step": 2800
+    },
+    {
+      "epoch": 0.84,
+      "eval_f1": 0.986085281999668,
+      "eval_f1_0": 0.9957410000522015,
+      "eval_f1_1": 0.9973185411943425,
+      "eval_f1_2": 0.9525071187726427,
+      "eval_f1_3": 0.9987744679794855,
+      "eval_loss": 0.025714803487062454,
+      "eval_runtime": 53.2058,
+      "eval_samples_per_second": 223.058,
+      "eval_steps_per_second": 0.883,
+      "step": 2800
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.5248607169136974e-05,
+      "loss": 0.0285,
+      "step": 2810
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.5222327341532645e-05,
+      "loss": 0.0372,
+      "step": 2820
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.519604751392831e-05,
+      "loss": 0.0675,
+      "step": 2830
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.5169767686323976e-05,
+      "loss": 0.0139,
+      "step": 2840
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.514348785871965e-05,
+      "loss": 0.0087,
+      "step": 2850
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.511720803111531e-05,
+      "loss": 0.0161,
+      "step": 2860
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.5090928203510984e-05,
+      "loss": 0.031,
+      "step": 2870
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.5064648375906656e-05,
+      "loss": 0.0776,
+      "step": 2880
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.503836854830232e-05,
+      "loss": 0.0099,
+      "step": 2890
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.501208872069799e-05,
+      "loss": 0.0671,
+      "step": 2900
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.4985808893093665e-05,
+      "loss": 0.0351,
+      "step": 2910
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.495952906548933e-05,
+      "loss": 0.0331,
+      "step": 2920
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.4933249237885e-05,
+      "loss": 0.0685,
+      "step": 2930
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.4906969410280666e-05,
+      "loss": 0.0241,
+      "step": 2940
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.488068958267634e-05,
+      "loss": 0.009,
+      "step": 2950
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.485440975507201e-05,
+      "loss": 0.0556,
+      "step": 2960
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.4828129927467675e-05,
+      "loss": 0.0149,
+      "step": 2970
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.480185009986335e-05,
+      "loss": 0.0253,
+      "step": 2980
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.477557027225902e-05,
+      "loss": 0.0231,
+      "step": 2990
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.4749290444654684e-05,
+      "loss": 0.0077,
+      "step": 3000
+    },
+    {
+      "epoch": 0.9,
+      "eval_f1": 0.9861304273680486,
+      "eval_f1_0": 0.9957041417243657,
+      "eval_f1_1": 0.995681458910857,
+      "eval_f1_2": 0.954885379973998,
+      "eval_f1_3": 0.9982507288629737,
+      "eval_loss": 0.030237887054681778,
+      "eval_runtime": 53.3094,
+      "eval_samples_per_second": 222.625,
+      "eval_steps_per_second": 0.882,
+      "step": 3000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.4723010617050355e-05,
+      "loss": 0.0344,
+      "step": 3010
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.469673078944602e-05,
+      "loss": 0.0149,
+      "step": 3020
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.467045096184169e-05,
+      "loss": 0.0027,
+      "step": 3030
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.4644171134237364e-05,
+      "loss": 0.0464,
+      "step": 3040
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.461789130663303e-05,
+      "loss": 0.0428,
+      "step": 3050
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.45916114790287e-05,
+      "loss": 0.0351,
+      "step": 3060
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.4565331651424366e-05,
+      "loss": 0.0128,
+      "step": 3070
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.453905182382004e-05,
+      "loss": 0.0105,
+      "step": 3080
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.451277199621571e-05,
+      "loss": 0.0478,
+      "step": 3090
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.4486492168611374e-05,
+      "loss": 0.0513,
+      "step": 3100
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.4460212341007046e-05,
+      "loss": 0.0271,
+      "step": 3110
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.443393251340272e-05,
+      "loss": 0.0166,
+      "step": 3120
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.440765268579838e-05,
+      "loss": 0.015,
+      "step": 3130
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.4381372858194055e-05,
+      "loss": 0.007,
+      "step": 3140
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.435509303058972e-05,
+      "loss": 0.0584,
+      "step": 3150
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.432881320298539e-05,
+      "loss": 0.0172,
+      "step": 3160
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.4302533375381063e-05,
+      "loss": 0.0437,
+      "step": 3170
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.427625354777673e-05,
+      "loss": 0.0336,
+      "step": 3180
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.42499737201724e-05,
+      "loss": 0.0203,
+      "step": 3190
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.422369389256807e-05,
+      "loss": 0.027,
+      "step": 3200
+    },
+    {
+      "epoch": 0.96,
+      "eval_f1": 0.9852472343292917,
+      "eval_f1_0": 0.9952913732973311,
+      "eval_f1_1": 0.997427579949319,
+      "eval_f1_2": 0.9496299706569509,
+      "eval_f1_3": 0.9986400134135663,
+      "eval_loss": 0.024080442264676094,
+      "eval_runtime": 53.231,
+      "eval_samples_per_second": 222.953,
+      "eval_steps_per_second": 0.883,
+      "step": 3200
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.419741406496374e-05,
+      "loss": 0.0407,
+      "step": 3210
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.417113423735941e-05,
+      "loss": 0.014,
+      "step": 3220
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.4144854409755074e-05,
+      "loss": 0.0247,
+      "step": 3230
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.4118574582150746e-05,
+      "loss": 0.0303,
+      "step": 3240
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.409229475454642e-05,
+      "loss": 0.0331,
+      "step": 3250
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.406601492694208e-05,
+      "loss": 0.0375,
+      "step": 3260
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.403973509933775e-05,
+      "loss": 0.0261,
+      "step": 3270
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.401345527173342e-05,
+      "loss": 0.0215,
+      "step": 3280
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.3987175444129084e-05,
+      "loss": 0.0314,
+      "step": 3290
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.3960895616524756e-05,
+      "loss": 0.0146,
+      "step": 3300
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.393461578892043e-05,
+      "loss": 0.0409,
+      "step": 3310
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.390833596131609e-05,
+      "loss": 0.0082,
+      "step": 3320
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.3882056133711765e-05,
+      "loss": 0.0079,
+      "step": 3330
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.385577630610743e-05,
+      "loss": 0.0265,
+      "step": 3340
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.38294964785031e-05,
+      "loss": 0.0169,
+      "step": 3350
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.380321665089877e-05,
+      "loss": 0.0257,
+      "step": 3360
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.377693682329444e-05,
+      "loss": 0.0228,
+      "step": 3370
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.375065699569011e-05,
+      "loss": 0.0229,
+      "step": 3380
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.3724377168085775e-05,
+      "loss": 0.0391,
+      "step": 3390
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.369809734048145e-05,
+      "loss": 0.044,
+      "step": 3400
+    },
+    {
+      "epoch": 1.02,
+      "eval_f1": 0.9704234487573644,
+      "eval_f1_0": 0.9898767933598651,
+      "eval_f1_1": 0.9964992059361791,
+      "eval_f1_2": 0.8965189093239978,
+      "eval_f1_3": 0.9987988864094153,
+      "eval_loss": 0.04618235304951668,
+      "eval_runtime": 53.2373,
+      "eval_samples_per_second": 222.926,
+      "eval_steps_per_second": 0.883,
+      "step": 3400
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.367181751287712e-05,
+      "loss": 0.0112,
+      "step": 3410
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.3645537685272784e-05,
+      "loss": 0.0091,
+      "step": 3420
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.3619257857668456e-05,
+      "loss": 0.0111,
+      "step": 3430
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.359297803006413e-05,
+      "loss": 0.0198,
+      "step": 3440
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.356669820245979e-05,
+      "loss": 0.0113,
+      "step": 3450
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.3540418374855464e-05,
+      "loss": 0.008,
+      "step": 3460
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.351413854725113e-05,
+      "loss": 0.0132,
+      "step": 3470
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.34878587196468e-05,
+      "loss": 0.0092,
+      "step": 3480
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.346157889204247e-05,
+      "loss": 0.0577,
+      "step": 3490
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.343529906443814e-05,
+      "loss": 0.0424,
+      "step": 3500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.340901923683381e-05,
+      "loss": 0.0248,
+      "step": 3510
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.338273940922948e-05,
+      "loss": 0.0133,
+      "step": 3520
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.3356459581625146e-05,
+      "loss": 0.0364,
+      "step": 3530
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.333017975402082e-05,
+      "loss": 0.0215,
+      "step": 3540
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.330389992641648e-05,
+      "loss": 0.0125,
+      "step": 3550
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.3277620098812155e-05,
+      "loss": 0.0247,
+      "step": 3560
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.325134027120783e-05,
+      "loss": 0.0077,
+      "step": 3570
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.322506044360349e-05,
+      "loss": 0.0237,
+      "step": 3580
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.3198780615999164e-05,
+      "loss": 0.0126,
+      "step": 3590
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.317250078839483e-05,
+      "loss": 0.0208,
+      "step": 3600
+    },
+    {
+      "epoch": 1.08,
+      "eval_f1": 0.9885796705027498,
+      "eval_f1_0": 0.996397910953713,
+      "eval_f1_1": 0.9976364871138669,
+      "eval_f1_2": 0.9614016994854276,
+      "eval_f1_3": 0.9988825844579914,
+      "eval_loss": 0.019902125000953674,
+      "eval_runtime": 53.2714,
+      "eval_samples_per_second": 222.784,
+      "eval_steps_per_second": 0.882,
+      "step": 3600
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.31462209607905e-05,
+      "loss": 0.0108,
+      "step": 3610
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.311994113318617e-05,
+      "loss": 0.0102,
+      "step": 3620
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.309366130558184e-05,
+      "loss": 0.0036,
+      "step": 3630
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.306738147797751e-05,
+      "loss": 0.0164,
+      "step": 3640
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.304110165037318e-05,
+      "loss": 0.0393,
+      "step": 3650
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.3014821822768846e-05,
+      "loss": 0.0416,
+      "step": 3660
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.298854199516452e-05,
+      "loss": 0.0322,
+      "step": 3670
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.296226216756018e-05,
+      "loss": 0.0081,
+      "step": 3680
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.293598233995585e-05,
+      "loss": 0.0171,
+      "step": 3690
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.290970251235152e-05,
+      "loss": 0.0094,
+      "step": 3700
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.2883422684747184e-05,
+      "loss": 0.0077,
+      "step": 3710
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.2857142857142856e-05,
+      "loss": 0.0139,
+      "step": 3720
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.283086302953853e-05,
+      "loss": 0.0382,
+      "step": 3730
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.280458320193419e-05,
+      "loss": 0.025,
+      "step": 3740
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.2778303374329865e-05,
+      "loss": 0.0164,
+      "step": 3750
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.275202354672554e-05,
+      "loss": 0.0062,
+      "step": 3760
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.27257437191212e-05,
+      "loss": 0.0452,
+      "step": 3770
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.2699463891516874e-05,
+      "loss": 0.0113,
+      "step": 3780
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.267318406391254e-05,
+      "loss": 0.0191,
+      "step": 3790
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.264690423630821e-05,
+      "loss": 0.0258,
+      "step": 3800
+    },
+    {
+      "epoch": 1.14,
+      "eval_f1": 0.9845698140822621,
+      "eval_f1_0": 0.9953415768896084,
+      "eval_f1_1": 0.9977729309531932,
+      "eval_f1_2": 0.9464030597896395,
+      "eval_f1_3": 0.9987616886966075,
+      "eval_loss": 0.030071619898080826,
+      "eval_runtime": 53.3053,
+      "eval_samples_per_second": 222.642,
+      "eval_steps_per_second": 0.882,
+      "step": 3800
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.262062440870388e-05,
+      "loss": 0.0106,
+      "step": 3810
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.259434458109955e-05,
+      "loss": 0.0339,
+      "step": 3820
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.256806475349522e-05,
+      "loss": 0.0177,
+      "step": 3830
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.254178492589089e-05,
+      "loss": 0.0256,
+      "step": 3840
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.2515505098286556e-05,
+      "loss": 0.013,
+      "step": 3850
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.248922527068223e-05,
+      "loss": 0.006,
+      "step": 3860
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.246294544307789e-05,
+      "loss": 0.0279,
+      "step": 3870
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2436665615473564e-05,
+      "loss": 0.032,
+      "step": 3880
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.2410385787869236e-05,
+      "loss": 0.0058,
+      "step": 3890
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.23841059602649e-05,
+      "loss": 0.0086,
+      "step": 3900
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.235782613266057e-05,
+      "loss": 0.0236,
+      "step": 3910
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.233154630505624e-05,
+      "loss": 0.0085,
+      "step": 3920
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.230526647745191e-05,
+      "loss": 0.0091,
+      "step": 3930
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.227898664984758e-05,
+      "loss": 0.0068,
+      "step": 3940
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.2252706822243247e-05,
+      "loss": 0.0123,
+      "step": 3950
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.222642699463892e-05,
+      "loss": 0.0034,
+      "step": 3960
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.220014716703459e-05,
+      "loss": 0.0324,
+      "step": 3970
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.2173867339430255e-05,
+      "loss": 0.0166,
+      "step": 3980
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.214758751182593e-05,
+      "loss": 0.02,
+      "step": 3990
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.212130768422159e-05,
+      "loss": 0.0243,
+      "step": 4000
+    },
+    {
+      "epoch": 1.2,
+      "eval_f1": 0.9758889262675604,
+      "eval_f1_0": 0.9918425590078939,
+      "eval_f1_1": 0.9975695123306997,
+      "eval_f1_2": 0.915422437863551,
+      "eval_f1_3": 0.9987211958680969,
+      "eval_loss": 0.05459127202630043,
+      "eval_runtime": 53.3211,
+      "eval_samples_per_second": 222.576,
+      "eval_steps_per_second": 0.881,
+      "step": 4000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.2095027856617264e-05,
+      "loss": 0.039,
+      "step": 4010
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.2068748029012936e-05,
+      "loss": 0.0142,
+      "step": 4020
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.20424682014086e-05,
+      "loss": 0.0372,
+      "step": 4030
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.201618837380427e-05,
+      "loss": 0.028,
+      "step": 4040
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.1989908546199944e-05,
+      "loss": 0.0376,
+      "step": 4050
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.196362871859561e-05,
+      "loss": 0.0241,
+      "step": 4060
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.193734889099128e-05,
+      "loss": 0.0144,
+      "step": 4070
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.1911069063386946e-05,
+      "loss": 0.0169,
+      "step": 4080
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.188478923578262e-05,
+      "loss": 0.0454,
+      "step": 4090
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.185850940817829e-05,
+      "loss": 0.0381,
+      "step": 4100
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.1832229580573955e-05,
+      "loss": 0.0355,
+      "step": 4110
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.180594975296962e-05,
+      "loss": 0.0217,
+      "step": 4120
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.177966992536529e-05,
+      "loss": 0.0134,
+      "step": 4130
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.1753390097760956e-05,
+      "loss": 0.0426,
+      "step": 4140
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.172711027015663e-05,
+      "loss": 0.0567,
+      "step": 4150
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.170083044255229e-05,
+      "loss": 0.0236,
+      "step": 4160
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.1674550614947965e-05,
+      "loss": 0.0253,
+      "step": 4170
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.164827078734364e-05,
+      "loss": 0.025,
+      "step": 4180
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.16219909597393e-05,
+      "loss": 0.0069,
+      "step": 4190
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.1595711132134974e-05,
+      "loss": 0.0315,
+      "step": 4200
+    },
+    {
+      "epoch": 1.26,
+      "eval_f1": 0.9905806489954309,
+      "eval_f1_0": 0.997055908346431,
+      "eval_f1_1": 0.997946052719898,
+      "eval_f1_2": 0.9692890107108816,
+      "eval_f1_3": 0.9980316242045127,
+      "eval_loss": 0.019053112715482712,
+      "eval_runtime": 53.2686,
+      "eval_samples_per_second": 222.796,
+      "eval_steps_per_second": 0.882,
+      "step": 4200
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.1569431304530645e-05,
+      "loss": 0.0309,
+      "step": 4210
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.154315147692631e-05,
+      "loss": 0.0152,
+      "step": 4220
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.151687164932198e-05,
+      "loss": 0.0236,
+      "step": 4230
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.149059182171765e-05,
+      "loss": 0.0144,
+      "step": 4240
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.146431199411332e-05,
+      "loss": 0.0243,
+      "step": 4250
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.143803216650899e-05,
+      "loss": 0.0146,
+      "step": 4260
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.1411752338904656e-05,
+      "loss": 0.0079,
+      "step": 4270
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.138547251130033e-05,
+      "loss": 0.0249,
+      "step": 4280
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.1359192683696e-05,
+      "loss": 0.0031,
+      "step": 4290
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.1332912856091665e-05,
+      "loss": 0.0055,
+      "step": 4300
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.1306633028487336e-05,
+      "loss": 0.0027,
+      "step": 4310
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.1280353200883e-05,
+      "loss": 0.0343,
+      "step": 4320
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.125407337327867e-05,
+      "loss": 0.0534,
+      "step": 4330
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.1227793545674345e-05,
+      "loss": 0.015,
+      "step": 4340
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.120151371807001e-05,
+      "loss": 0.0439,
+      "step": 4350
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.117523389046568e-05,
+      "loss": 0.0127,
+      "step": 4360
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.114895406286135e-05,
+      "loss": 0.0342,
+      "step": 4370
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.112267423525702e-05,
+      "loss": 0.037,
+      "step": 4380
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.109639440765269e-05,
+      "loss": 0.0419,
+      "step": 4390
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.1070114580048355e-05,
+      "loss": 0.0107,
+      "step": 4400
+    },
+    {
+      "epoch": 1.32,
+      "eval_f1": 0.9901851558278552,
+      "eval_f1_0": 0.9968996202993475,
+      "eval_f1_1": 0.9974876010233711,
+      "eval_f1_2": 0.967474022334325,
+      "eval_f1_3": 0.9988793796543771,
+      "eval_loss": 0.016864441335201263,
+      "eval_runtime": 53.2892,
+      "eval_samples_per_second": 222.709,
+      "eval_steps_per_second": 0.882,
+      "step": 4400
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.104383475244403e-05,
+      "loss": 0.0278,
+      "step": 4410
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.10175549248397e-05,
+      "loss": 0.0091,
+      "step": 4420
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.0991275097235364e-05,
+      "loss": 0.0264,
+      "step": 4430
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.0964995269631036e-05,
+      "loss": 0.0202,
+      "step": 4440
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.09387154420267e-05,
+      "loss": 0.0267,
+      "step": 4450
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.091243561442237e-05,
+      "loss": 0.0211,
+      "step": 4460
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.0886155786818044e-05,
+      "loss": 0.0127,
+      "step": 4470
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.085987595921371e-05,
+      "loss": 0.0134,
+      "step": 4480
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.083359613160938e-05,
+      "loss": 0.0136,
+      "step": 4490
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.080731630400505e-05,
+      "loss": 0.0079,
+      "step": 4500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.078103647640072e-05,
+      "loss": 0.0278,
+      "step": 4510
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.075475664879639e-05,
+      "loss": 0.0109,
+      "step": 4520
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.0728476821192055e-05,
+      "loss": 0.003,
+      "step": 4530
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.070219699358773e-05,
+      "loss": 0.0403,
+      "step": 4540
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.067591716598339e-05,
+      "loss": 0.0177,
+      "step": 4550
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.064963733837906e-05,
+      "loss": 0.0049,
+      "step": 4560
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.062335751077473e-05,
+      "loss": 0.0047,
+      "step": 4570
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.05970776831704e-05,
+      "loss": 0.0455,
+      "step": 4580
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.0570797855566065e-05,
+      "loss": 0.0087,
+      "step": 4590
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.054451802796174e-05,
+      "loss": 0.0077,
+      "step": 4600
+    },
+    {
+      "epoch": 1.38,
+      "eval_f1": 0.9810789471197044,
+      "eval_f1_0": 0.9943578765323313,
+      "eval_f1_1": 0.9973014848536501,
+      "eval_f1_2": 0.9337456942425872,
+      "eval_f1_3": 0.998910732850249,
+      "eval_loss": 0.029722465202212334,
+      "eval_runtime": 53.2615,
+      "eval_samples_per_second": 222.825,
+      "eval_steps_per_second": 0.882,
+      "step": 4600
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.051823820035741e-05,
+      "loss": 0.0127,
+      "step": 4610
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.0491958372753074e-05,
+      "loss": 0.0151,
+      "step": 4620
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.0465678545148746e-05,
+      "loss": 0.004,
+      "step": 4630
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.043939871754441e-05,
+      "loss": 0.0059,
+      "step": 4640
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.041311888994008e-05,
+      "loss": 0.004,
+      "step": 4650
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.0386839062335754e-05,
+      "loss": 0.0211,
+      "step": 4660
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.036055923473142e-05,
+      "loss": 0.025,
+      "step": 4670
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.033427940712709e-05,
+      "loss": 0.0145,
+      "step": 4680
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.0307999579522756e-05,
+      "loss": 0.0226,
+      "step": 4690
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.028171975191843e-05,
+      "loss": 0.0197,
+      "step": 4700
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.02554399243141e-05,
+      "loss": 0.0217,
+      "step": 4710
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.0229160096709765e-05,
+      "loss": 0.019,
+      "step": 4720
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.0202880269105437e-05,
+      "loss": 0.049,
+      "step": 4730
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.017660044150111e-05,
+      "loss": 0.0171,
+      "step": 4740
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.015032061389677e-05,
+      "loss": 0.0113,
+      "step": 4750
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.0124040786292445e-05,
+      "loss": 0.0374,
+      "step": 4760
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.009776095868811e-05,
+      "loss": 0.0166,
+      "step": 4770
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.007148113108378e-05,
+      "loss": 0.0092,
+      "step": 4780
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.0045201303479454e-05,
+      "loss": 0.0268,
+      "step": 4790
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.001892147587512e-05,
+      "loss": 0.0303,
+      "step": 4800
+    },
+    {
+      "epoch": 1.44,
+      "eval_f1": 0.9882341464888815,
+      "eval_f1_0": 0.9962570940403399,
+      "eval_f1_1": 0.9961376820738735,
+      "eval_f1_2": 0.9616341565869263,
+      "eval_f1_3": 0.9989076532543865,
+      "eval_loss": 0.022203443571925163,
+      "eval_runtime": 53.3025,
+      "eval_samples_per_second": 222.654,
+      "eval_steps_per_second": 0.882,
+      "step": 4800
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.999264164827079e-05,
+      "loss": 0.0113,
+      "step": 4810
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.996636182066646e-05,
+      "loss": 0.0202,
+      "step": 4820
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.994008199306213e-05,
+      "loss": 0.0292,
+      "step": 4830
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.99138021654578e-05,
+      "loss": 0.0194,
+      "step": 4840
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.9887522337853464e-05,
+      "loss": 0.0077,
+      "step": 4850
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.9861242510249136e-05,
+      "loss": 0.0038,
+      "step": 4860
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.983496268264481e-05,
+      "loss": 0.0252,
+      "step": 4870
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.980868285504047e-05,
+      "loss": 0.0037,
+      "step": 4880
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.9782403027436145e-05,
+      "loss": 0.0271,
+      "step": 4890
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.975612319983181e-05,
+      "loss": 0.0082,
+      "step": 4900
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.972984337222748e-05,
+      "loss": 0.0241,
+      "step": 4910
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.970356354462315e-05,
+      "loss": 0.024,
+      "step": 4920
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.967728371701882e-05,
+      "loss": 0.037,
+      "step": 4930
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.965100388941449e-05,
+      "loss": 0.009,
+      "step": 4940
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.962472406181016e-05,
+      "loss": 0.0031,
+      "step": 4950
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.959844423420583e-05,
+      "loss": 0.0348,
+      "step": 4960
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.957216440660149e-05,
+      "loss": 0.0059,
+      "step": 4970
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.9545884578997164e-05,
+      "loss": 0.0187,
+      "step": 4980
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.951960475139283e-05,
+      "loss": 0.028,
+      "step": 4990
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.94933249237885e-05,
+      "loss": 0.0062,
+      "step": 5000
+    },
+    {
+      "epoch": 1.5,
+      "eval_f1": 0.988427145836447,
+      "eval_f1_0": 0.9964087925447377,
+      "eval_f1_1": 0.9971002048830786,
+      "eval_f1_2": 0.9613856561134559,
+      "eval_f1_3": 0.9988139298045158,
+      "eval_loss": 0.019996481016278267,
+      "eval_runtime": 53.2705,
+      "eval_samples_per_second": 222.787,
+      "eval_steps_per_second": 0.882,
+      "step": 5000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.9467045096184165e-05,
+      "loss": 0.0114,
+      "step": 5010
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.944076526857984e-05,
+      "loss": 0.0051,
+      "step": 5020
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.941448544097551e-05,
+      "loss": 0.0106,
+      "step": 5030
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.9388205613371174e-05,
+      "loss": 0.0079,
+      "step": 5040
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.9361925785766846e-05,
+      "loss": 0.0106,
+      "step": 5050
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.933564595816252e-05,
+      "loss": 0.0048,
+      "step": 5060
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.930936613055818e-05,
+      "loss": 0.0113,
+      "step": 5070
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.9283086302953854e-05,
+      "loss": 0.0108,
+      "step": 5080
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.925680647534952e-05,
+      "loss": 0.0289,
+      "step": 5090
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.923052664774519e-05,
+      "loss": 0.0267,
+      "step": 5100
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.920424682014086e-05,
+      "loss": 0.0063,
+      "step": 5110
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.917796699253653e-05,
+      "loss": 0.0094,
+      "step": 5120
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.91516871649322e-05,
+      "loss": 0.0061,
+      "step": 5130
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.912540733732787e-05,
+      "loss": 0.0369,
+      "step": 5140
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.909912750972354e-05,
+      "loss": 0.0138,
+      "step": 5150
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.907284768211921e-05,
+      "loss": 0.0172,
+      "step": 5160
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.9046567854514874e-05,
+      "loss": 0.0069,
+      "step": 5170
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.9020288026910545e-05,
+      "loss": 0.0148,
+      "step": 5180
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.899400819930622e-05,
+      "loss": 0.006,
+      "step": 5190
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.896772837170188e-05,
+      "loss": 0.0244,
+      "step": 5200
+    },
+    {
+      "epoch": 1.56,
+      "eval_f1": 0.9863014050948063,
+      "eval_f1_0": 0.9958320662985854,
+      "eval_f1_1": 0.9981034555423159,
+      "eval_f1_2": 0.9524213321580693,
+      "eval_f1_3": 0.9988487663802547,
+      "eval_loss": 0.02705535478889942,
+      "eval_runtime": 53.231,
+      "eval_samples_per_second": 222.953,
+      "eval_steps_per_second": 0.883,
+      "step": 5200
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.8941448544097554e-05,
+      "loss": 0.0424,
+      "step": 5210
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.891516871649322e-05,
+      "loss": 0.0091,
+      "step": 5220
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.888888888888889e-05,
+      "loss": 0.01,
+      "step": 5230
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.886260906128456e-05,
+      "loss": 0.0246,
+      "step": 5240
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.883632923368023e-05,
+      "loss": 0.0163,
+      "step": 5250
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.88100494060759e-05,
+      "loss": 0.007,
+      "step": 5260
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.878376957847157e-05,
+      "loss": 0.0041,
+      "step": 5270
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.8757489750867236e-05,
+      "loss": 0.0206,
+      "step": 5280
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.873120992326291e-05,
+      "loss": 0.0287,
+      "step": 5290
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.870493009565857e-05,
+      "loss": 0.0163,
+      "step": 5300
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.8678650268054245e-05,
+      "loss": 0.0217,
+      "step": 5310
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.8652370440449917e-05,
+      "loss": 0.0282,
+      "step": 5320
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.862609061284558e-05,
+      "loss": 0.0044,
+      "step": 5330
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.8599810785241253e-05,
+      "loss": 0.016,
+      "step": 5340
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.8573530957636925e-05,
+      "loss": 0.005,
+      "step": 5350
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.854725113003259e-05,
+      "loss": 0.0182,
+      "step": 5360
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.852097130242826e-05,
+      "loss": 0.0237,
+      "step": 5370
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.849469147482393e-05,
+      "loss": 0.0266,
+      "step": 5380
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.84684116472196e-05,
+      "loss": 0.0332,
+      "step": 5390
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.8442131819615264e-05,
+      "loss": 0.0123,
+      "step": 5400
+    },
+    {
+      "epoch": 1.62,
+      "eval_f1": 0.9903904030593307,
+      "eval_f1_0": 0.9969964432107575,
+      "eval_f1_1": 0.9975871590904739,
+      "eval_f1_2": 0.9680795194695785,
+      "eval_f1_3": 0.9988984904665126,
+      "eval_loss": 0.016976816579699516,
+      "eval_runtime": 53.3087,
+      "eval_samples_per_second": 222.628,
+      "eval_steps_per_second": 0.882,
+      "step": 5400
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.841585199201093e-05,
+      "loss": 0.0089,
+      "step": 5410
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.83895721644066e-05,
+      "loss": 0.0202,
+      "step": 5420
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.836329233680227e-05,
+      "loss": 0.0339,
+      "step": 5430
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.833701250919794e-05,
+      "loss": 0.0285,
+      "step": 5440
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.831073268159361e-05,
+      "loss": 0.0101,
+      "step": 5450
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.828445285398928e-05,
+      "loss": 0.0385,
+      "step": 5460
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.8258173026384946e-05,
+      "loss": 0.0088,
+      "step": 5470
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.823189319878062e-05,
+      "loss": 0.0192,
+      "step": 5480
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.820561337117628e-05,
+      "loss": 0.0084,
+      "step": 5490
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.8179333543571955e-05,
+      "loss": 0.0431,
+      "step": 5500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.8153053715967626e-05,
+      "loss": 0.0348,
+      "step": 5510
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.812677388836329e-05,
+      "loss": 0.019,
+      "step": 5520
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.810049406075896e-05,
+      "loss": 0.054,
+      "step": 5530
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.807421423315463e-05,
+      "loss": 0.0313,
+      "step": 5540
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.80479344055503e-05,
+      "loss": 0.0209,
+      "step": 5550
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.802165457794597e-05,
+      "loss": 0.0187,
+      "step": 5560
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.799537475034164e-05,
+      "loss": 0.0177,
+      "step": 5570
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.796909492273731e-05,
+      "loss": 0.0175,
+      "step": 5580
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.794281509513298e-05,
+      "loss": 0.0075,
+      "step": 5590
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.7916535267528646e-05,
+      "loss": 0.0428,
+      "step": 5600
+    },
+    {
+      "epoch": 1.68,
+      "eval_f1": 0.9895366266198364,
+      "eval_f1_0": 0.9966985214279319,
+      "eval_f1_1": 0.9971627586656717,
+      "eval_f1_2": 0.9655738501665956,
+      "eval_f1_3": 0.9987113762191467,
+      "eval_loss": 0.017527282238006592,
+      "eval_runtime": 53.24,
+      "eval_samples_per_second": 222.915,
+      "eval_steps_per_second": 0.883,
+      "step": 5600
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.789025543992432e-05,
+      "loss": 0.0396,
+      "step": 5610
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.786397561231998e-05,
+      "loss": 0.0139,
+      "step": 5620
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.7837695784715654e-05,
+      "loss": 0.0222,
+      "step": 5630
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.7811415957111326e-05,
+      "loss": 0.0363,
+      "step": 5640
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.778513612950699e-05,
+      "loss": 0.0023,
+      "step": 5650
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.775885630190266e-05,
+      "loss": 0.053,
+      "step": 5660
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.7732576474298335e-05,
+      "loss": 0.009,
+      "step": 5670
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.7706296646694e-05,
+      "loss": 0.0065,
+      "step": 5680
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.768001681908967e-05,
+      "loss": 0.032,
+      "step": 5690
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.7653736991485336e-05,
+      "loss": 0.0351,
+      "step": 5700
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.762745716388101e-05,
+      "loss": 0.0168,
+      "step": 5710
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.760117733627668e-05,
+      "loss": 0.0149,
+      "step": 5720
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.7574897508672345e-05,
+      "loss": 0.0114,
+      "step": 5730
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.754861768106802e-05,
+      "loss": 0.0131,
+      "step": 5740
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.752233785346368e-05,
+      "loss": 0.0143,
+      "step": 5750
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.7496058025859354e-05,
+      "loss": 0.0167,
+      "step": 5760
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.7469778198255025e-05,
+      "loss": 0.027,
+      "step": 5770
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.744349837065069e-05,
+      "loss": 0.0247,
+      "step": 5780
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.741721854304636e-05,
+      "loss": 0.012,
+      "step": 5790
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.7390938715442034e-05,
+      "loss": 0.0144,
+      "step": 5800
+    },
+    {
+      "epoch": 1.74,
+      "eval_f1": 0.9902148365969656,
+      "eval_f1_0": 0.9968994883788248,
+      "eval_f1_1": 0.997476188470126,
+      "eval_f1_2": 0.968282018778334,
+      "eval_f1_3": 0.9982016507605777,
+      "eval_loss": 0.016549725085496902,
+      "eval_runtime": 53.2744,
+      "eval_samples_per_second": 222.771,
+      "eval_steps_per_second": 0.882,
+      "step": 5800
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.73646588878377e-05,
+      "loss": 0.0109,
+      "step": 5810
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.733837906023337e-05,
+      "loss": 0.0063,
+      "step": 5820
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.7312099232629036e-05,
+      "loss": 0.0207,
+      "step": 5830
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.72858194050247e-05,
+      "loss": 0.0088,
+      "step": 5840
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.725953957742037e-05,
+      "loss": 0.0513,
+      "step": 5850
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.723325974981604e-05,
+      "loss": 0.0525,
+      "step": 5860
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.720697992221171e-05,
+      "loss": 0.0101,
+      "step": 5870
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.718070009460738e-05,
+      "loss": 0.0406,
+      "step": 5880
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.7154420267003046e-05,
+      "loss": 0.0115,
+      "step": 5890
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.712814043939872e-05,
+      "loss": 0.0174,
+      "step": 5900
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.710186061179439e-05,
+      "loss": 0.0199,
+      "step": 5910
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.7075580784190055e-05,
+      "loss": 0.0089,
+      "step": 5920
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.704930095658573e-05,
+      "loss": 0.0172,
+      "step": 5930
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.702302112898139e-05,
+      "loss": 0.0235,
+      "step": 5940
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.6996741301377063e-05,
+      "loss": 0.0181,
+      "step": 5950
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.6970461473772735e-05,
+      "loss": 0.0254,
+      "step": 5960
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.69441816461684e-05,
+      "loss": 0.0115,
+      "step": 5970
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.691790181856407e-05,
+      "loss": 0.0048,
+      "step": 5980
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.689162199095974e-05,
+      "loss": 0.0312,
+      "step": 5990
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.686534216335541e-05,
+      "loss": 0.032,
+      "step": 6000
+    },
+    {
+      "epoch": 1.8,
+      "eval_f1": 0.990454296038497,
+      "eval_f1_0": 0.9969953781583025,
+      "eval_f1_1": 0.9976073527889294,
+      "eval_f1_2": 0.9686538913295297,
+      "eval_f1_3": 0.9985605618772259,
+      "eval_loss": 0.01473264116793871,
+      "eval_runtime": 53.276,
+      "eval_samples_per_second": 222.764,
+      "eval_steps_per_second": 0.882,
+      "step": 6000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.683906233575108e-05,
+      "loss": 0.0064,
+      "step": 6010
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.6812782508146746e-05,
+      "loss": 0.0431,
+      "step": 6020
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.678650268054242e-05,
+      "loss": 0.0349,
+      "step": 6030
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.676022285293809e-05,
+      "loss": 0.0249,
+      "step": 6040
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.6733943025333754e-05,
+      "loss": 0.0143,
+      "step": 6050
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.6707663197729426e-05,
+      "loss": 0.0292,
+      "step": 6060
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.668138337012509e-05,
+      "loss": 0.0099,
+      "step": 6070
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.665510354252076e-05,
+      "loss": 0.0101,
+      "step": 6080
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.6628823714916435e-05,
+      "loss": 0.0081,
+      "step": 6090
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.66025438873121e-05,
+      "loss": 0.0236,
+      "step": 6100
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.657626405970777e-05,
+      "loss": 0.0386,
+      "step": 6110
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.654998423210344e-05,
+      "loss": 0.0038,
+      "step": 6120
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.652370440449911e-05,
+      "loss": 0.0098,
+      "step": 6130
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.649742457689478e-05,
+      "loss": 0.011,
+      "step": 6140
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.6471144749290445e-05,
+      "loss": 0.0199,
+      "step": 6150
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.644486492168612e-05,
+      "loss": 0.0083,
+      "step": 6160
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.641858509408179e-05,
+      "loss": 0.0516,
+      "step": 6170
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.6392305266477454e-05,
+      "loss": 0.0181,
+      "step": 6180
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.6366025438873126e-05,
+      "loss": 0.0135,
+      "step": 6190
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.633974561126879e-05,
+      "loss": 0.0062,
+      "step": 6200
+    },
+    {
+      "epoch": 1.86,
+      "eval_f1": 0.9828945374124921,
+      "eval_f1_0": 0.9944139820295181,
+      "eval_f1_1": 0.9976678621430353,
+      "eval_f1_2": 0.9408426096553008,
+      "eval_f1_3": 0.9986536958221142,
+      "eval_loss": 0.030880145728588104,
+      "eval_runtime": 53.2137,
+      "eval_samples_per_second": 223.025,
+      "eval_steps_per_second": 0.883,
+      "step": 6200
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.631346578366446e-05,
+      "loss": 0.0288,
+      "step": 6210
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.6287185956060134e-05,
+      "loss": 0.017,
+      "step": 6220
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.62609061284558e-05,
+      "loss": 0.0222,
+      "step": 6230
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.623462630085147e-05,
+      "loss": 0.0193,
+      "step": 6240
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.6208346473247136e-05,
+      "loss": 0.007,
+      "step": 6250
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.61820666456428e-05,
+      "loss": 0.014,
+      "step": 6260
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.615578681803847e-05,
+      "loss": 0.0117,
+      "step": 6270
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.6129506990434145e-05,
+      "loss": 0.0147,
+      "step": 6280
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.610322716282981e-05,
+      "loss": 0.0028,
+      "step": 6290
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.607694733522548e-05,
+      "loss": 0.0376,
+      "step": 6300
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.6050667507621146e-05,
+      "loss": 0.0256,
+      "step": 6310
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.602438768001682e-05,
+      "loss": 0.0355,
+      "step": 6320
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.599810785241249e-05,
+      "loss": 0.0048,
+      "step": 6330
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.5971828024808155e-05,
+      "loss": 0.0036,
+      "step": 6340
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.594554819720383e-05,
+      "loss": 0.0072,
+      "step": 6350
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.59192683695995e-05,
+      "loss": 0.002,
+      "step": 6360
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.5892988541995164e-05,
+      "loss": 0.0043,
+      "step": 6370
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.5866708714390835e-05,
+      "loss": 0.0231,
+      "step": 6380
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.58404288867865e-05,
+      "loss": 0.0042,
+      "step": 6390
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.581414905918217e-05,
+      "loss": 0.0131,
+      "step": 6400
+    },
+    {
+      "epoch": 1.92,
+      "eval_f1": 0.9884566672672931,
+      "eval_f1_0": 0.9963622833229593,
+      "eval_f1_1": 0.9980315149284771,
+      "eval_f1_2": 0.9604415328973178,
+      "eval_f1_3": 0.9989913379204182,
+      "eval_loss": 0.022728843614459038,
+      "eval_runtime": 53.2888,
+      "eval_samples_per_second": 222.711,
+      "eval_steps_per_second": 0.882,
+      "step": 6400
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.5787869231577844e-05,
+      "loss": 0.0559,
+      "step": 6410
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.576158940397351e-05,
+      "loss": 0.0036,
+      "step": 6420
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.573530957636918e-05,
+      "loss": 0.0274,
+      "step": 6430
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.570902974876485e-05,
+      "loss": 0.0248,
+      "step": 6440
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.568274992116052e-05,
+      "loss": 0.0219,
+      "step": 6450
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.565647009355619e-05,
+      "loss": 0.013,
+      "step": 6460
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.5630190265951855e-05,
+      "loss": 0.0144,
+      "step": 6470
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.5603910438347526e-05,
+      "loss": 0.0284,
+      "step": 6480
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.55776306107432e-05,
+      "loss": 0.0067,
+      "step": 6490
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.555135078313886e-05,
+      "loss": 0.0123,
+      "step": 6500
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.5525070955534535e-05,
+      "loss": 0.0276,
+      "step": 6510
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.54987911279302e-05,
+      "loss": 0.0761,
+      "step": 6520
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.547251130032587e-05,
+      "loss": 0.0257,
+      "step": 6530
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.5446231472721544e-05,
+      "loss": 0.0064,
+      "step": 6540
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.541995164511721e-05,
+      "loss": 0.0103,
+      "step": 6550
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.539367181751288e-05,
+      "loss": 0.0039,
+      "step": 6560
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.536739198990855e-05,
+      "loss": 0.0103,
+      "step": 6570
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.534111216230422e-05,
+      "loss": 0.0119,
+      "step": 6580
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.531483233469989e-05,
+      "loss": 0.0133,
+      "step": 6590
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.5288552507095554e-05,
+      "loss": 0.0376,
+      "step": 6600
+    },
+    {
+      "epoch": 1.98,
+      "eval_f1": 0.993526599442064,
+      "eval_f1_0": 0.9979671388907684,
+      "eval_f1_1": 0.9974921281221278,
+      "eval_f1_2": 0.9797800483751776,
+      "eval_f1_3": 0.9988670823801823,
+      "eval_loss": 0.013585735112428665,
+      "eval_runtime": 53.2249,
+      "eval_samples_per_second": 222.978,
+      "eval_steps_per_second": 0.883,
+      "step": 6600
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.5262272679491226e-05,
+      "loss": 0.0107,
+      "step": 6610
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.52359928518869e-05,
+      "loss": 0.0116,
+      "step": 6620
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.520971302428256e-05,
+      "loss": 0.0099,
+      "step": 6630
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.5183433196678234e-05,
+      "loss": 0.0122,
+      "step": 6640
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.5157153369073906e-05,
+      "loss": 0.0059,
+      "step": 6650
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.513087354146957e-05,
+      "loss": 0.0271,
+      "step": 6660
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.510459371386524e-05,
+      "loss": 0.004,
+      "step": 6670
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.507831388626091e-05,
+      "loss": 0.0079,
+      "step": 6680
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.505203405865657e-05,
+      "loss": 0.0117,
+      "step": 6690
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 3.5025754231052245e-05,
+      "loss": 0.007,
+      "step": 6700
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 3.499947440344791e-05,
+      "loss": 0.0051,
+      "step": 6710
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 3.497319457584358e-05,
+      "loss": 0.0025,
+      "step": 6720
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 3.4946914748239253e-05,
+      "loss": 0.0085,
+      "step": 6730
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 3.492063492063492e-05,
+      "loss": 0.0136,
+      "step": 6740
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 3.489435509303059e-05,
+      "loss": 0.0093,
+      "step": 6750
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.486807526542626e-05,
+      "loss": 0.0068,
+      "step": 6760
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.484179543782193e-05,
+      "loss": 0.0195,
+      "step": 6770
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.48155156102176e-05,
+      "loss": 0.0055,
+      "step": 6780
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.4789235782613264e-05,
+      "loss": 0.0026,
+      "step": 6790
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.4762955955008936e-05,
+      "loss": 0.0057,
+      "step": 6800
+    },
+    {
+      "epoch": 2.04,
+      "eval_f1": 0.9890607674618193,
+      "eval_f1_0": 0.9964958601655933,
+      "eval_f1_1": 0.998043157856405,
+      "eval_f1_2": 0.9627095323146126,
+      "eval_f1_3": 0.9989945195106662,
+      "eval_loss": 0.022696251049637794,
+      "eval_runtime": 53.3706,
+      "eval_samples_per_second": 222.37,
+      "eval_steps_per_second": 0.881,
+      "step": 6800
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.473667612740461e-05,
+      "loss": 0.0086,
+      "step": 6810
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.471039629980027e-05,
+      "loss": 0.0041,
+      "step": 6820
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 3.4684116472195944e-05,
+      "loss": 0.0027,
+      "step": 6830
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 3.465783664459161e-05,
+      "loss": 0.0026,
+      "step": 6840
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 3.463155681698728e-05,
+      "loss": 0.0076,
+      "step": 6850
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 3.460527698938295e-05,
+      "loss": 0.007,
+      "step": 6860
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 3.457899716177862e-05,
+      "loss": 0.007,
+      "step": 6870
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 3.455271733417429e-05,
+      "loss": 0.0106,
+      "step": 6880
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 3.452643750656996e-05,
+      "loss": 0.0392,
+      "step": 6890
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 3.4500157678965626e-05,
+      "loss": 0.0118,
+      "step": 6900
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 3.44738778513613e-05,
+      "loss": 0.0076,
+      "step": 6910
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 3.444759802375696e-05,
+      "loss": 0.0029,
+      "step": 6920
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.4421318196152635e-05,
+      "loss": 0.0113,
+      "step": 6930
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.439503836854831e-05,
+      "loss": 0.0094,
+      "step": 6940
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.436875854094397e-05,
+      "loss": 0.0056,
+      "step": 6950
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 3.4342478713339644e-05,
+      "loss": 0.0026,
+      "step": 6960
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 3.4316198885735316e-05,
+      "loss": 0.0025,
+      "step": 6970
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 3.428991905813098e-05,
+      "loss": 0.002,
+      "step": 6980
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 3.426363923052665e-05,
+      "loss": 0.0022,
+      "step": 6990
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 3.423735940292232e-05,
+      "loss": 0.034,
+      "step": 7000
+    },
+    {
+      "epoch": 2.1,
+      "eval_f1": 0.9919433062214094,
+      "eval_f1_0": 0.9974867382429036,
+      "eval_f1_1": 0.9978065802592223,
+      "eval_f1_2": 0.973924993088714,
+      "eval_f1_3": 0.9985549132947977,
+      "eval_loss": 0.016368065029382706,
+      "eval_runtime": 53.1976,
+      "eval_samples_per_second": 223.093,
+      "eval_steps_per_second": 0.883,
+      "step": 7000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 3.421107957531799e-05,
+      "loss": 0.0051,
+      "step": 7010
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 3.418479974771366e-05,
+      "loss": 0.0312,
+      "step": 7020
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 3.4158519920109326e-05,
+      "loss": 0.0208,
+      "step": 7030
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 3.4132240092505e-05,
+      "loss": 0.0029,
+      "step": 7040
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 3.410596026490066e-05,
+      "loss": 0.0051,
+      "step": 7050
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.4079680437296335e-05,
+      "loss": 0.0193,
+      "step": 7060
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.4053400609692006e-05,
+      "loss": 0.0017,
+      "step": 7070
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.402712078208767e-05,
+      "loss": 0.0066,
+      "step": 7080
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.400084095448334e-05,
+      "loss": 0.0093,
+      "step": 7090
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 3.3974561126879015e-05,
+      "loss": 0.0198,
+      "step": 7100
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 3.394828129927467e-05,
+      "loss": 0.0127,
+      "step": 7110
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 3.3922001471670345e-05,
+      "loss": 0.0038,
+      "step": 7120
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.389572164406602e-05,
+      "loss": 0.0057,
+      "step": 7130
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.386944181646168e-05,
+      "loss": 0.0129,
+      "step": 7140
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.3843161988857354e-05,
+      "loss": 0.0208,
+      "step": 7150
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.381688216125302e-05,
+      "loss": 0.0145,
+      "step": 7160
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.379060233364869e-05,
+      "loss": 0.0099,
+      "step": 7170
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.376432250604436e-05,
+      "loss": 0.0067,
+      "step": 7180
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.373804267844003e-05,
+      "loss": 0.0206,
+      "step": 7190
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.37117628508357e-05,
+      "loss": 0.0048,
+      "step": 7200
+    },
+    {
+      "epoch": 2.16,
+      "eval_f1": 0.9929448755536868,
+      "eval_f1_0": 0.9977800469610748,
+      "eval_f1_1": 0.9977676687864864,
+      "eval_f1_2": 0.9772373793014063,
+      "eval_f1_3": 0.99899440716578,
+      "eval_loss": 0.013158726505935192,
+      "eval_runtime": 53.2712,
+      "eval_samples_per_second": 222.785,
+      "eval_steps_per_second": 0.882,
+      "step": 7200
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.368548302323137e-05,
+      "loss": 0.0036,
+      "step": 7210
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.3659203195627036e-05,
+      "loss": 0.0071,
+      "step": 7220
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.363292336802271e-05,
+      "loss": 0.0334,
+      "step": 7230
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.360664354041837e-05,
+      "loss": 0.0117,
+      "step": 7240
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.3580363712814044e-05,
+      "loss": 0.0033,
+      "step": 7250
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.3554083885209716e-05,
+      "loss": 0.0229,
+      "step": 7260
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.352780405760538e-05,
+      "loss": 0.0112,
+      "step": 7270
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.350152423000105e-05,
+      "loss": 0.0052,
+      "step": 7280
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.3475244402396725e-05,
+      "loss": 0.0048,
+      "step": 7290
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 3.344896457479239e-05,
+      "loss": 0.0059,
+      "step": 7300
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 3.342268474718806e-05,
+      "loss": 0.0227,
+      "step": 7310
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 3.339640491958373e-05,
+      "loss": 0.0036,
+      "step": 7320
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.33701250919794e-05,
+      "loss": 0.0157,
+      "step": 7330
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.334384526437507e-05,
+      "loss": 0.01,
+      "step": 7340
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.3317565436770735e-05,
+      "loss": 0.0048,
+      "step": 7350
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.329128560916641e-05,
+      "loss": 0.0025,
+      "step": 7360
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 3.326500578156207e-05,
+      "loss": 0.0375,
+      "step": 7370
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 3.3238725953957744e-05,
+      "loss": 0.0029,
+      "step": 7380
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 3.3212446126353416e-05,
+      "loss": 0.0149,
+      "step": 7390
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 3.318616629874908e-05,
+      "loss": 0.0039,
+      "step": 7400
+    },
+    {
+      "epoch": 2.22,
+      "eval_f1": 0.9916341644454706,
+      "eval_f1_0": 0.9974010783434527,
+      "eval_f1_1": 0.9975840038042353,
+      "eval_f1_2": 0.9729829725969001,
+      "eval_f1_3": 0.998568603037294,
+      "eval_loss": 0.01573426090180874,
+      "eval_runtime": 53.3004,
+      "eval_samples_per_second": 222.662,
+      "eval_steps_per_second": 0.882,
+      "step": 7400
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 3.315988647114475e-05,
+      "loss": 0.0061,
+      "step": 7410
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 3.3133606643540424e-05,
+      "loss": 0.0081,
+      "step": 7420
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 3.310732681593609e-05,
+      "loss": 0.0066,
+      "step": 7430
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 3.308104698833176e-05,
+      "loss": 0.0151,
+      "step": 7440
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 3.3054767160727426e-05,
+      "loss": 0.0083,
+      "step": 7450
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 3.30284873331231e-05,
+      "loss": 0.0202,
+      "step": 7460
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.300220750551877e-05,
+      "loss": 0.0239,
+      "step": 7470
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.2975927677914435e-05,
+      "loss": 0.0037,
+      "step": 7480
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.2949647850310107e-05,
+      "loss": 0.0192,
+      "step": 7490
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 3.292336802270578e-05,
+      "loss": 0.0313,
+      "step": 7500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 3.289708819510144e-05,
+      "loss": 0.0158,
+      "step": 7510
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 3.2870808367497115e-05,
+      "loss": 0.009,
+      "step": 7520
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 3.284452853989278e-05,
+      "loss": 0.0101,
+      "step": 7530
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 3.2818248712288445e-05,
+      "loss": 0.0055,
+      "step": 7540
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 3.279196888468412e-05,
+      "loss": 0.006,
+      "step": 7550
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 3.276568905707978e-05,
+      "loss": 0.0142,
+      "step": 7560
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.2739409229475454e-05,
+      "loss": 0.0434,
+      "step": 7570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.2713129401871126e-05,
+      "loss": 0.0362,
+      "step": 7580
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.268684957426679e-05,
+      "loss": 0.011,
+      "step": 7590
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 3.266056974666246e-05,
+      "loss": 0.007,
+      "step": 7600
+    },
+    {
+      "epoch": 2.28,
+      "eval_f1": 0.9892171960228088,
+      "eval_f1_0": 0.9965462294958259,
+      "eval_f1_1": 0.9981181971485513,
+      "eval_f1_2": 0.9632160634438539,
+      "eval_f1_3": 0.998988294003004,
+      "eval_loss": 0.01945377327501774,
+      "eval_runtime": 53.2216,
+      "eval_samples_per_second": 222.992,
+      "eval_steps_per_second": 0.883,
+      "step": 7600
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 3.2634289919058134e-05,
+      "loss": 0.0128,
+      "step": 7610
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 3.26080100914538e-05,
+      "loss": 0.0029,
+      "step": 7620
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 3.258173026384947e-05,
+      "loss": 0.0025,
+      "step": 7630
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 3.2555450436245136e-05,
+      "loss": 0.0301,
+      "step": 7640
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 3.252917060864081e-05,
+      "loss": 0.0211,
+      "step": 7650
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 3.250289078103648e-05,
+      "loss": 0.0018,
+      "step": 7660
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 3.2476610953432145e-05,
+      "loss": 0.0047,
+      "step": 7670
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 3.2450331125827816e-05,
+      "loss": 0.0255,
+      "step": 7680
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 3.242405129822348e-05,
+      "loss": 0.0257,
+      "step": 7690
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.239777147061915e-05,
+      "loss": 0.01,
+      "step": 7700
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.2371491643014825e-05,
+      "loss": 0.004,
+      "step": 7710
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.234521181541049e-05,
+      "loss": 0.0159,
+      "step": 7720
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 3.231893198780616e-05,
+      "loss": 0.011,
+      "step": 7730
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 3.2292652160201834e-05,
+      "loss": 0.0043,
+      "step": 7740
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 3.22663723325975e-05,
+      "loss": 0.0137,
+      "step": 7750
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 3.224009250499317e-05,
+      "loss": 0.0147,
+      "step": 7760
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 3.2213812677388835e-05,
+      "loss": 0.0112,
+      "step": 7770
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 3.218753284978451e-05,
+      "loss": 0.028,
+      "step": 7780
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 3.216125302218018e-05,
+      "loss": 0.0033,
+      "step": 7790
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 3.2134973194575844e-05,
+      "loss": 0.0038,
+      "step": 7800
+    },
+    {
+      "epoch": 2.34,
+      "eval_f1": 0.9918897161538254,
+      "eval_f1_0": 0.9974320930508085,
+      "eval_f1_1": 0.9974823323244807,
+      "eval_f1_2": 0.973631544785807,
+      "eval_f1_3": 0.9990128944542052,
+      "eval_loss": 0.014562987722456455,
+      "eval_runtime": 53.2607,
+      "eval_samples_per_second": 222.829,
+      "eval_steps_per_second": 0.882,
+      "step": 7800
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 3.2108693366971516e-05,
+      "loss": 0.0061,
+      "step": 7810
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 3.208241353936719e-05,
+      "loss": 0.0074,
+      "step": 7820
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.205613371176285e-05,
+      "loss": 0.0038,
+      "step": 7830
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.2029853884158525e-05,
+      "loss": 0.0479,
+      "step": 7840
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.200357405655419e-05,
+      "loss": 0.004,
+      "step": 7850
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.197729422894986e-05,
+      "loss": 0.0063,
+      "step": 7860
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 3.195101440134553e-05,
+      "loss": 0.0077,
+      "step": 7870
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 3.19247345737412e-05,
+      "loss": 0.0057,
+      "step": 7880
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 3.189845474613687e-05,
+      "loss": 0.0138,
+      "step": 7890
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 3.1872174918532535e-05,
+      "loss": 0.0082,
+      "step": 7900
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 3.184589509092821e-05,
+      "loss": 0.0029,
+      "step": 7910
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 3.181961526332388e-05,
+      "loss": 0.0063,
+      "step": 7920
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 3.1793335435719544e-05,
+      "loss": 0.0097,
+      "step": 7930
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 3.1767055608115215e-05,
+      "loss": 0.0068,
+      "step": 7940
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 3.174077578051089e-05,
+      "loss": 0.0042,
+      "step": 7950
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 3.1714495952906545e-05,
+      "loss": 0.0115,
+      "step": 7960
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.168821612530222e-05,
+      "loss": 0.008,
+      "step": 7970
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.166193629769789e-05,
+      "loss": 0.014,
+      "step": 7980
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.1635656470093554e-05,
+      "loss": 0.0083,
+      "step": 7990
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 3.1609376642489226e-05,
+      "loss": 0.0364,
+      "step": 8000
+    },
+    {
+      "epoch": 2.4,
+      "eval_f1": 0.9886284900276143,
+      "eval_f1_0": 0.9963441156646601,
+      "eval_f1_1": 0.9983078229174053,
+      "eval_f1_2": 0.9608833388103845,
+      "eval_f1_3": 0.9989786827180078,
+      "eval_loss": 0.019674289971590042,
+      "eval_runtime": 53.2331,
+      "eval_samples_per_second": 222.944,
+      "eval_steps_per_second": 0.883,
+      "step": 8000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 3.158309681488489e-05,
+      "loss": 0.0109,
+      "step": 8010
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 3.155681698728056e-05,
+      "loss": 0.0038,
+      "step": 8020
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.1530537159676234e-05,
+      "loss": 0.0111,
+      "step": 8030
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.15042573320719e-05,
+      "loss": 0.0054,
+      "step": 8040
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.147797750446757e-05,
+      "loss": 0.0022,
+      "step": 8050
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.145169767686324e-05,
+      "loss": 0.0193,
+      "step": 8060
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.142541784925891e-05,
+      "loss": 0.0041,
+      "step": 8070
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.139913802165458e-05,
+      "loss": 0.0039,
+      "step": 8080
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.1372858194050245e-05,
+      "loss": 0.0022,
+      "step": 8090
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.134657836644592e-05,
+      "loss": 0.0043,
+      "step": 8100
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.132029853884159e-05,
+      "loss": 0.0013,
+      "step": 8110
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.1294018711237253e-05,
+      "loss": 0.0074,
+      "step": 8120
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.1267738883632925e-05,
+      "loss": 0.0187,
+      "step": 8130
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.124145905602859e-05,
+      "loss": 0.0279,
+      "step": 8140
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.121517922842426e-05,
+      "loss": 0.0279,
+      "step": 8150
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.1188899400819934e-05,
+      "loss": 0.0239,
+      "step": 8160
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.11626195732156e-05,
+      "loss": 0.0121,
+      "step": 8170
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.113633974561127e-05,
+      "loss": 0.0039,
+      "step": 8180
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.111005991800694e-05,
+      "loss": 0.0031,
+      "step": 8190
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.108378009040261e-05,
+      "loss": 0.0024,
+      "step": 8200
+    },
+    {
+      "epoch": 2.46,
+      "eval_f1": 0.9887433509677206,
+      "eval_f1_0": 0.9963817531063979,
+      "eval_f1_1": 0.9976038783023411,
+      "eval_f1_2": 0.9621265373211634,
+      "eval_f1_3": 0.9988612351409804,
+      "eval_loss": 0.021771090105175972,
+      "eval_runtime": 53.2034,
+      "eval_samples_per_second": 223.069,
+      "eval_steps_per_second": 0.883,
+      "step": 8200
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.105750026279828e-05,
+      "loss": 0.0024,
+      "step": 8210
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.1031220435193944e-05,
+      "loss": 0.0189,
+      "step": 8220
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.1004940607589616e-05,
+      "loss": 0.0027,
+      "step": 8230
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.097866077998529e-05,
+      "loss": 0.0056,
+      "step": 8240
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.095238095238095e-05,
+      "loss": 0.0026,
+      "step": 8250
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.0926101124776625e-05,
+      "loss": 0.0143,
+      "step": 8260
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.0899821297172296e-05,
+      "loss": 0.0223,
+      "step": 8270
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.087354146956796e-05,
+      "loss": 0.0057,
+      "step": 8280
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.084726164196363e-05,
+      "loss": 0.0109,
+      "step": 8290
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.08209818143593e-05,
+      "loss": 0.0019,
+      "step": 8300
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.079470198675497e-05,
+      "loss": 0.0026,
+      "step": 8310
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.076842215915064e-05,
+      "loss": 0.0338,
+      "step": 8320
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.074214233154631e-05,
+      "loss": 0.0014,
+      "step": 8330
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.071586250394198e-05,
+      "loss": 0.0018,
+      "step": 8340
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.0689582676337644e-05,
+      "loss": 0.0016,
+      "step": 8350
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.0663302848733316e-05,
+      "loss": 0.001,
+      "step": 8360
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.063702302112899e-05,
+      "loss": 0.0039,
+      "step": 8370
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.061074319352465e-05,
+      "loss": 0.0014,
+      "step": 8380
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.058446336592032e-05,
+      "loss": 0.0062,
+      "step": 8390
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.055818353831599e-05,
+      "loss": 0.0134,
+      "step": 8400
+    },
+    {
+      "epoch": 2.52,
+      "eval_f1": 0.9928095103361599,
+      "eval_f1_0": 0.9977114497558153,
+      "eval_f1_1": 0.9976035367962546,
+      "eval_f1_2": 0.9769068149909359,
+      "eval_f1_3": 0.9990162398016342,
+      "eval_loss": 0.017637787386775017,
+      "eval_runtime": 53.3768,
+      "eval_samples_per_second": 222.344,
+      "eval_steps_per_second": 0.881,
+      "step": 8400
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.0531903710711654e-05,
+      "loss": 0.0014,
+      "step": 8410
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.0505623883107326e-05,
+      "loss": 0.0336,
+      "step": 8420
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0479344055502994e-05,
+      "loss": 0.0107,
+      "step": 8430
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0453064227898663e-05,
+      "loss": 0.016,
+      "step": 8440
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0426784400294335e-05,
+      "loss": 0.0037,
+      "step": 8450
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0400504572690003e-05,
+      "loss": 0.0154,
+      "step": 8460
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.037422474508567e-05,
+      "loss": 0.0022,
+      "step": 8470
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.034794491748134e-05,
+      "loss": 0.0081,
+      "step": 8480
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.032166508987701e-05,
+      "loss": 0.0065,
+      "step": 8490
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.029538526227268e-05,
+      "loss": 0.0083,
+      "step": 8500
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.026910543466835e-05,
+      "loss": 0.0056,
+      "step": 8510
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.0242825607064017e-05,
+      "loss": 0.0091,
+      "step": 8520
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.021654577945969e-05,
+      "loss": 0.0632,
+      "step": 8530
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.0190265951855357e-05,
+      "loss": 0.0235,
+      "step": 8540
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.0163986124251025e-05,
+      "loss": 0.0332,
+      "step": 8550
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.0137706296646694e-05,
+      "loss": 0.0091,
+      "step": 8560
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.0111426469042366e-05,
+      "loss": 0.0224,
+      "step": 8570
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.0085146641438034e-05,
+      "loss": 0.0069,
+      "step": 8580
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.0058866813833702e-05,
+      "loss": 0.0048,
+      "step": 8590
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 3.003258698622937e-05,
+      "loss": 0.0038,
+      "step": 8600
+    },
+    {
+      "epoch": 2.58,
+      "eval_f1": 0.992581772092419,
+      "eval_f1_0": 0.9976360293746089,
+      "eval_f1_1": 0.9976559952429364,
+      "eval_f1_2": 0.9759631250352142,
+      "eval_f1_3": 0.9990719387169165,
+      "eval_loss": 0.01393364742398262,
+      "eval_runtime": 53.2808,
+      "eval_samples_per_second": 222.745,
+      "eval_steps_per_second": 0.882,
+      "step": 8600
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 3.0006307158625043e-05,
+      "loss": 0.0254,
+      "step": 8610
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.998002733102071e-05,
+      "loss": 0.0098,
+      "step": 8620
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.995374750341638e-05,
+      "loss": 0.0032,
+      "step": 8630
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.9927467675812048e-05,
+      "loss": 0.0347,
+      "step": 8640
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.9901187848207716e-05,
+      "loss": 0.0425,
+      "step": 8650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.9874908020603388e-05,
+      "loss": 0.0113,
+      "step": 8660
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.9848628192999056e-05,
+      "loss": 0.0147,
+      "step": 8670
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.9822348365394725e-05,
+      "loss": 0.0162,
+      "step": 8680
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.9796068537790393e-05,
+      "loss": 0.003,
+      "step": 8690
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.9769788710186065e-05,
+      "loss": 0.0043,
+      "step": 8700
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.9743508882581733e-05,
+      "loss": 0.0419,
+      "step": 8710
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.9717229054977402e-05,
+      "loss": 0.0136,
+      "step": 8720
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.969094922737307e-05,
+      "loss": 0.0033,
+      "step": 8730
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.9664669399768742e-05,
+      "loss": 0.0092,
+      "step": 8740
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.963838957216441e-05,
+      "loss": 0.0058,
+      "step": 8750
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.961210974456008e-05,
+      "loss": 0.0106,
+      "step": 8760
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.9585829916955747e-05,
+      "loss": 0.0117,
+      "step": 8770
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.955955008935142e-05,
+      "loss": 0.0127,
+      "step": 8780
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.9533270261747088e-05,
+      "loss": 0.0265,
+      "step": 8790
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.9506990434142756e-05,
+      "loss": 0.0076,
+      "step": 8800
+    },
+    {
+      "epoch": 2.64,
+      "eval_f1": 0.9915600607688022,
+      "eval_f1_0": 0.99730177462593,
+      "eval_f1_1": 0.9973903465849999,
+      "eval_f1_2": 0.9726622709368042,
+      "eval_f1_3": 0.9988858509274748,
+      "eval_loss": 0.015386177226901054,
+      "eval_runtime": 53.1887,
+      "eval_samples_per_second": 223.13,
+      "eval_steps_per_second": 0.884,
+      "step": 8800
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.948071060653842e-05,
+      "loss": 0.0312,
+      "step": 8810
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.945443077893409e-05,
+      "loss": 0.0111,
+      "step": 8820
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.9428150951329758e-05,
+      "loss": 0.0195,
+      "step": 8830
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.9401871123725426e-05,
+      "loss": 0.0067,
+      "step": 8840
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.9375591296121098e-05,
+      "loss": 0.0074,
+      "step": 8850
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.9349311468516766e-05,
+      "loss": 0.0023,
+      "step": 8860
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.9323031640912435e-05,
+      "loss": 0.0057,
+      "step": 8870
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.9296751813308103e-05,
+      "loss": 0.0023,
+      "step": 8880
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.9270471985703775e-05,
+      "loss": 0.0024,
+      "step": 8890
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.9244192158099443e-05,
+      "loss": 0.0092,
+      "step": 8900
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.9217912330495112e-05,
+      "loss": 0.0227,
+      "step": 8910
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.919163250289078e-05,
+      "loss": 0.0227,
+      "step": 8920
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.9165352675286452e-05,
+      "loss": 0.0139,
+      "step": 8930
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.913907284768212e-05,
+      "loss": 0.0138,
+      "step": 8940
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.911279302007779e-05,
+      "loss": 0.0035,
+      "step": 8950
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.9086513192473457e-05,
+      "loss": 0.003,
+      "step": 8960
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.9060233364869126e-05,
+      "loss": 0.0088,
+      "step": 8970
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.9033953537264797e-05,
+      "loss": 0.0132,
+      "step": 8980
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.9007673709660466e-05,
+      "loss": 0.025,
+      "step": 8990
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.8981393882056134e-05,
+      "loss": 0.0059,
+      "step": 9000
+    },
+    {
+      "epoch": 2.7,
+      "eval_f1": 0.9888104160718681,
+      "eval_f1_0": 0.996429967486629,
+      "eval_f1_1": 0.9980465417654738,
+      "eval_f1_2": 0.9619597795354117,
+      "eval_f1_3": 0.9988053754999581,
+      "eval_loss": 0.021114500239491463,
+      "eval_runtime": 53.2111,
+      "eval_samples_per_second": 223.036,
+      "eval_steps_per_second": 0.883,
+      "step": 9000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.8955114054451803e-05,
+      "loss": 0.0153,
+      "step": 9010
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.8928834226847474e-05,
+      "loss": 0.021,
+      "step": 9020
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.8902554399243143e-05,
+      "loss": 0.0072,
+      "step": 9030
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.887627457163881e-05,
+      "loss": 0.0058,
+      "step": 9040
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.884999474403448e-05,
+      "loss": 0.0048,
+      "step": 9050
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.882371491643015e-05,
+      "loss": 0.0198,
+      "step": 9060
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.879743508882582e-05,
+      "loss": 0.0062,
+      "step": 9070
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.8771155261221488e-05,
+      "loss": 0.0024,
+      "step": 9080
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.8744875433617157e-05,
+      "loss": 0.012,
+      "step": 9090
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.871859560601283e-05,
+      "loss": 0.0047,
+      "step": 9100
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.8692315778408497e-05,
+      "loss": 0.0029,
+      "step": 9110
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.8666035950804165e-05,
+      "loss": 0.0566,
+      "step": 9120
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.8639756123199834e-05,
+      "loss": 0.0027,
+      "step": 9130
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.8613476295595505e-05,
+      "loss": 0.0017,
+      "step": 9140
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.8587196467991174e-05,
+      "loss": 0.0018,
+      "step": 9150
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.8560916640386842e-05,
+      "loss": 0.0016,
+      "step": 9160
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.853463681278251e-05,
+      "loss": 0.0131,
+      "step": 9170
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.850835698517818e-05,
+      "loss": 0.0104,
+      "step": 9180
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 2.848207715757385e-05,
+      "loss": 0.0016,
+      "step": 9190
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.845579732996952e-05,
+      "loss": 0.0143,
+      "step": 9200
+    },
+    {
+      "epoch": 2.76,
+      "eval_f1": 0.992309982102097,
+      "eval_f1_0": 0.9975932931756409,
+      "eval_f1_1": 0.9982769954941229,
+      "eval_f1_2": 0.9744498013904024,
+      "eval_f1_3": 0.9989198383482217,
+      "eval_loss": 0.013883223757147789,
+      "eval_runtime": 53.1516,
+      "eval_samples_per_second": 223.286,
+      "eval_steps_per_second": 0.884,
+      "step": 9200
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.8429517502365188e-05,
+      "loss": 0.0019,
+      "step": 9210
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 2.8403237674760856e-05,
+      "loss": 0.0101,
+      "step": 9220
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.8376957847156528e-05,
+      "loss": 0.002,
+      "step": 9230
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.835067801955219e-05,
+      "loss": 0.0161,
+      "step": 9240
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.8324398191947858e-05,
+      "loss": 0.0182,
+      "step": 9250
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 2.829811836434353e-05,
+      "loss": 0.0153,
+      "step": 9260
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.8271838536739198e-05,
+      "loss": 0.0102,
+      "step": 9270
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.8245558709134867e-05,
+      "loss": 0.0021,
+      "step": 9280
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 2.8219278881530535e-05,
+      "loss": 0.0038,
+      "step": 9290
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 2.8192999053926207e-05,
+      "loss": 0.0119,
+      "step": 9300
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 2.8166719226321875e-05,
+      "loss": 0.0234,
+      "step": 9310
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 2.8140439398717544e-05,
+      "loss": 0.0144,
+      "step": 9320
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.8114159571113212e-05,
+      "loss": 0.0058,
+      "step": 9330
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.8087879743508884e-05,
+      "loss": 0.0023,
+      "step": 9340
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.8061599915904552e-05,
+      "loss": 0.006,
+      "step": 9350
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 2.803532008830022e-05,
+      "loss": 0.0049,
+      "step": 9360
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 2.800904026069589e-05,
+      "loss": 0.0062,
+      "step": 9370
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 2.798276043309156e-05,
+      "loss": 0.0052,
+      "step": 9380
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 2.795648060548723e-05,
+      "loss": 0.0271,
+      "step": 9390
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.7930200777882898e-05,
+      "loss": 0.025,
+      "step": 9400
+    },
+    {
+      "epoch": 2.82,
+      "eval_f1": 0.9851190515011226,
+      "eval_f1_0": 0.9951271173259356,
+      "eval_f1_1": 0.997643322972136,
+      "eval_f1_2": 0.9486835194800827,
+      "eval_f1_3": 0.9990222462263361,
+      "eval_loss": 0.028859291225671768,
+      "eval_runtime": 53.2539,
+      "eval_samples_per_second": 222.857,
+      "eval_steps_per_second": 0.883,
+      "step": 9400
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.7903920950278566e-05,
+      "loss": 0.0183,
+      "step": 9410
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 2.7877641122674238e-05,
+      "loss": 0.0018,
+      "step": 9420
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.7851361295069906e-05,
+      "loss": 0.0129,
+      "step": 9430
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.7825081467465575e-05,
+      "loss": 0.002,
+      "step": 9440
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.7798801639861243e-05,
+      "loss": 0.0016,
+      "step": 9450
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 2.777252181225691e-05,
+      "loss": 0.004,
+      "step": 9460
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 2.7746241984652583e-05,
+      "loss": 0.018,
+      "step": 9470
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 2.771996215704825e-05,
+      "loss": 0.0204,
+      "step": 9480
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 2.769368232944392e-05,
+      "loss": 0.0038,
+      "step": 9490
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.766740250183959e-05,
+      "loss": 0.0214,
+      "step": 9500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.764112267423526e-05,
+      "loss": 0.0058,
+      "step": 9510
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.761484284663093e-05,
+      "loss": 0.009,
+      "step": 9520
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7588563019026597e-05,
+      "loss": 0.0037,
+      "step": 9530
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7562283191422265e-05,
+      "loss": 0.0013,
+      "step": 9540
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7536003363817937e-05,
+      "loss": 0.0104,
+      "step": 9550
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7509723536213606e-05,
+      "loss": 0.0013,
+      "step": 9560
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.7483443708609274e-05,
+      "loss": 0.0029,
+      "step": 9570
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.7457163881004942e-05,
+      "loss": 0.0227,
+      "step": 9580
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.7430884053400614e-05,
+      "loss": 0.008,
+      "step": 9590
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.7404604225796283e-05,
+      "loss": 0.0036,
+      "step": 9600
+    },
+    {
+      "epoch": 2.88,
+      "eval_f1": 0.9923914591031513,
+      "eval_f1_0": 0.997616570752836,
+      "eval_f1_1": 0.9979790000076698,
+      "eval_f1_2": 0.9749075484011311,
+      "eval_f1_3": 0.9990627172509683,
+      "eval_loss": 0.013414908200502396,
+      "eval_runtime": 53.3061,
+      "eval_samples_per_second": 222.639,
+      "eval_steps_per_second": 0.882,
+      "step": 9600
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.737832439819195e-05,
+      "loss": 0.002,
+      "step": 9610
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.735204457058762e-05,
+      "loss": 0.0192,
+      "step": 9620
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.732576474298329e-05,
+      "loss": 0.003,
+      "step": 9630
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.729948491537896e-05,
+      "loss": 0.0056,
+      "step": 9640
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.7273205087774628e-05,
+      "loss": 0.0165,
+      "step": 9650
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.7246925260170297e-05,
+      "loss": 0.0099,
+      "step": 9660
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.722064543256596e-05,
+      "loss": 0.0075,
+      "step": 9670
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.719436560496163e-05,
+      "loss": 0.0019,
+      "step": 9680
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.71680857773573e-05,
+      "loss": 0.0021,
+      "step": 9690
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.714180594975297e-05,
+      "loss": 0.0109,
+      "step": 9700
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.711552612214864e-05,
+      "loss": 0.023,
+      "step": 9710
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.7089246294544307e-05,
+      "loss": 0.0027,
+      "step": 9720
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.7062966466939975e-05,
+      "loss": 0.0035,
+      "step": 9730
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.7036686639335647e-05,
+      "loss": 0.0036,
+      "step": 9740
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.7010406811731316e-05,
+      "loss": 0.0024,
+      "step": 9750
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.6984126984126984e-05,
+      "loss": 0.0049,
+      "step": 9760
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 2.6957847156522652e-05,
+      "loss": 0.0185,
+      "step": 9770
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 2.693156732891832e-05,
+      "loss": 0.0057,
+      "step": 9780
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 2.6905287501313993e-05,
+      "loss": 0.0021,
+      "step": 9790
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.687900767370966e-05,
+      "loss": 0.0028,
+      "step": 9800
+    },
+    {
+      "epoch": 2.94,
+      "eval_f1": 0.9918819873108211,
+      "eval_f1_0": 0.9974759191679455,
+      "eval_f1_1": 0.9981394867589323,
+      "eval_f1_2": 0.9730079967192946,
+      "eval_f1_3": 0.9989045465971121,
+      "eval_loss": 0.01633334718644619,
+      "eval_runtime": 53.2283,
+      "eval_samples_per_second": 222.964,
+      "eval_steps_per_second": 0.883,
+      "step": 9800
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.685272784610533e-05,
+      "loss": 0.008,
+      "step": 9810
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.6826448018500998e-05,
+      "loss": 0.0211,
+      "step": 9820
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 2.680016819089667e-05,
+      "loss": 0.0201,
+      "step": 9830
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.6773888363292338e-05,
+      "loss": 0.0162,
+      "step": 9840
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.6747608535688006e-05,
+      "loss": 0.0107,
+      "step": 9850
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.6721328708083675e-05,
+      "loss": 0.0047,
+      "step": 9860
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.6695048880479347e-05,
+      "loss": 0.0052,
+      "step": 9870
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.6668769052875015e-05,
+      "loss": 0.0069,
+      "step": 9880
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.6642489225270683e-05,
+      "loss": 0.0029,
+      "step": 9890
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.6616209397666352e-05,
+      "loss": 0.0019,
+      "step": 9900
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.6589929570062024e-05,
+      "loss": 0.0098,
+      "step": 9910
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.6563649742457692e-05,
+      "loss": 0.0133,
+      "step": 9920
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.653736991485336e-05,
+      "loss": 0.0141,
+      "step": 9930
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.651109008724903e-05,
+      "loss": 0.0761,
+      "step": 9940
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.64848102596447e-05,
+      "loss": 0.0017,
+      "step": 9950
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.645853043204037e-05,
+      "loss": 0.0135,
+      "step": 9960
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.6432250604436037e-05,
+      "loss": 0.0088,
+      "step": 9970
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.6405970776831706e-05,
+      "loss": 0.0071,
+      "step": 9980
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.6379690949227374e-05,
+      "loss": 0.0033,
+      "step": 9990
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.6353411121623046e-05,
+      "loss": 0.0038,
+      "step": 10000
+    },
+    {
+      "epoch": 3.0,
+      "eval_f1": 0.9837313140464833,
+      "eval_f1_0": 0.9951107926131856,
+      "eval_f1_1": 0.9979246154406312,
+      "eval_f1_2": 0.9429821948777298,
+      "eval_f1_3": 0.9989076532543865,
+      "eval_loss": 0.04059078171849251,
+      "eval_runtime": 53.2583,
+      "eval_samples_per_second": 222.839,
+      "eval_steps_per_second": 0.882,
+      "step": 10000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.6327131294018714e-05,
+      "loss": 0.0319,
+      "step": 10010
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.6300851466414383e-05,
+      "loss": 0.0336,
+      "step": 10020
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.627457163881005e-05,
+      "loss": 0.0063,
+      "step": 10030
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 2.6248291811205723e-05,
+      "loss": 0.009,
+      "step": 10040
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 2.622201198360139e-05,
+      "loss": 0.0084,
+      "step": 10050
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 2.619573215599706e-05,
+      "loss": 0.0014,
+      "step": 10060
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 2.6169452328392728e-05,
+      "loss": 0.0065,
+      "step": 10070
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 2.61431725007884e-05,
+      "loss": 0.0013,
+      "step": 10080
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 2.6116892673184062e-05,
+      "loss": 0.0048,
+      "step": 10090
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 2.609061284557973e-05,
+      "loss": 0.0073,
+      "step": 10100
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 2.6064333017975402e-05,
+      "loss": 0.0082,
+      "step": 10110
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 2.603805319037107e-05,
+      "loss": 0.0014,
+      "step": 10120
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 2.601177336276674e-05,
+      "loss": 0.0076,
+      "step": 10130
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 2.5985493535162407e-05,
+      "loss": 0.0092,
+      "step": 10140
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 2.595921370755808e-05,
+      "loss": 0.021,
+      "step": 10150
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 2.5932933879953747e-05,
+      "loss": 0.0036,
+      "step": 10160
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 2.5906654052349416e-05,
+      "loss": 0.0022,
+      "step": 10170
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 2.5880374224745084e-05,
+      "loss": 0.0199,
+      "step": 10180
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 2.5854094397140756e-05,
+      "loss": 0.0033,
+      "step": 10190
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 2.5827814569536424e-05,
+      "loss": 0.0121,
+      "step": 10200
+    },
+    {
+      "epoch": 3.06,
+      "eval_f1": 0.9884805553259847,
+      "eval_f1_0": 0.9963046652881733,
+      "eval_f1_1": 0.9979486825557499,
+      "eval_f1_2": 0.960680592016051,
+      "eval_f1_3": 0.9989882814439645,
+      "eval_loss": 0.02228163555264473,
+      "eval_runtime": 53.3261,
+      "eval_samples_per_second": 222.555,
+      "eval_steps_per_second": 0.881,
+      "step": 10200
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 2.5801534741932093e-05,
+      "loss": 0.0025,
+      "step": 10210
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 2.577525491432776e-05,
+      "loss": 0.0011,
+      "step": 10220
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 2.5748975086723433e-05,
+      "loss": 0.0033,
+      "step": 10230
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 2.57226952591191e-05,
+      "loss": 0.0029,
+      "step": 10240
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 2.569641543151477e-05,
+      "loss": 0.0118,
+      "step": 10250
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 2.5670135603910438e-05,
+      "loss": 0.0021,
+      "step": 10260
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 2.5643855776306107e-05,
+      "loss": 0.0129,
+      "step": 10270
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 2.561757594870178e-05,
+      "loss": 0.0025,
+      "step": 10280
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 2.5591296121097447e-05,
+      "loss": 0.0318,
+      "step": 10290
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 2.5565016293493115e-05,
+      "loss": 0.0046,
+      "step": 10300
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 2.5538736465888784e-05,
+      "loss": 0.0038,
+      "step": 10310
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 2.5512456638284455e-05,
+      "loss": 0.0114,
+      "step": 10320
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 2.5486176810680124e-05,
+      "loss": 0.0024,
+      "step": 10330
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 2.5459896983075792e-05,
+      "loss": 0.0026,
+      "step": 10340
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 2.543361715547146e-05,
+      "loss": 0.0019,
+      "step": 10350
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 2.5407337327867132e-05,
+      "loss": 0.0057,
+      "step": 10360
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 2.53810575002628e-05,
+      "loss": 0.0012,
+      "step": 10370
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 2.535477767265847e-05,
+      "loss": 0.0109,
+      "step": 10380
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 2.5328497845054138e-05,
+      "loss": 0.0036,
+      "step": 10390
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 2.530221801744981e-05,
+      "loss": 0.0029,
+      "step": 10400
+    },
+    {
+      "epoch": 3.12,
+      "eval_f1": 0.9915827182707018,
+      "eval_f1_0": 0.9973477738948522,
+      "eval_f1_1": 0.9979634260466687,
+      "eval_f1_2": 0.9723783772229823,
+      "eval_f1_3": 0.9986412959183041,
+      "eval_loss": 0.019460231065750122,
+      "eval_runtime": 53.2808,
+      "eval_samples_per_second": 222.744,
+      "eval_steps_per_second": 0.882,
+      "step": 10400
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 2.5275938189845478e-05,
+      "loss": 0.0022,
+      "step": 10410
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 2.5249658362241146e-05,
+      "loss": 0.0035,
+      "step": 10420
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 2.5223378534636815e-05,
+      "loss": 0.0017,
+      "step": 10430
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 2.5197098707032486e-05,
+      "loss": 0.005,
+      "step": 10440
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 2.5170818879428155e-05,
+      "loss": 0.003,
+      "step": 10450
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 2.5144539051823823e-05,
+      "loss": 0.02,
+      "step": 10460
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 2.511825922421949e-05,
+      "loss": 0.0036,
+      "step": 10470
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 2.509197939661516e-05,
+      "loss": 0.0013,
+      "step": 10480
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 2.5065699569010832e-05,
+      "loss": 0.0015,
+      "step": 10490
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2.50394197414065e-05,
+      "loss": 0.0028,
+      "step": 10500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2.501313991380217e-05,
+      "loss": 0.0026,
+      "step": 10510
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2.4986860086197837e-05,
+      "loss": 0.0037,
+      "step": 10520
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2.4960580258593506e-05,
+      "loss": 0.0564,
+      "step": 10530
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 2.4934300430989174e-05,
+      "loss": 0.0096,
+      "step": 10540
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 2.4908020603384842e-05,
+      "loss": 0.0016,
+      "step": 10550
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 2.4881740775780514e-05,
+      "loss": 0.0052,
+      "step": 10560
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 2.4855460948176183e-05,
+      "loss": 0.0012,
+      "step": 10570
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 2.482918112057185e-05,
+      "loss": 0.0031,
+      "step": 10580
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 2.480290129296752e-05,
+      "loss": 0.0028,
+      "step": 10590
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 2.477662146536319e-05,
+      "loss": 0.0019,
+      "step": 10600
+    },
+    {
+      "epoch": 3.18,
+      "eval_f1": 0.9911705258446226,
+      "eval_f1_0": 0.9972116420434848,
+      "eval_f1_1": 0.9983713230222113,
+      "eval_f1_2": 0.9701202907562665,
+      "eval_f1_3": 0.9989788475565281,
+      "eval_loss": 0.02063840441405773,
+      "eval_runtime": 53.403,
+      "eval_samples_per_second": 222.235,
+      "eval_steps_per_second": 0.88,
+      "step": 10600
+    },
+    {
+      "epoch": 3.18,
+      "step": 10600,
+      "total_flos": 2.1233219239359283e+17,
+      "train_loss": 0.03157895615913044,
+      "train_runtime": 7385.7528,
+      "train_samples_per_second": 86.764,
+      "train_steps_per_second": 2.712
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 20028,
+  "num_train_epochs": 6,
+  "save_steps": 200,
+  "total_flos": 2.1233219239359283e+17,
+  "trial_name": null,
+  "trial_params": null
+}