diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,6673 @@
+{
+  "best_metric": 0.9506666666666667,
+  "best_model_checkpoint": "ResNet_AI_image_detector/checkpoint-10930",
+  "epoch": 9.993142857142857,
+  "global_step": 10930,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.574565416285453e-07,
+      "loss": 0.6931,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.149130832570906e-07,
+      "loss": 0.6973,
+      "step": 20
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.372369624885636e-06,
+      "loss": 0.6927,
+      "step": 30
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.8298261665141813e-06,
+      "loss": 0.692,
+      "step": 40
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.287282708142727e-06,
+      "loss": 0.6921,
+      "step": 50
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.744739249771272e-06,
+      "loss": 0.6885,
+      "step": 60
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.2021957913998174e-06,
+      "loss": 0.6869,
+      "step": 70
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.6596523330283626e-06,
+      "loss": 0.6886,
+      "step": 80
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.1171088746569085e-06,
+      "loss": 0.6932,
+      "step": 90
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.574565416285454e-06,
+      "loss": 0.6853,
+      "step": 100
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.032021957913998e-06,
+      "loss": 0.6879,
+      "step": 110
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.489478499542544e-06,
+      "loss": 0.6854,
+      "step": 120
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.946935041171089e-06,
+      "loss": 0.688,
+      "step": 130
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 6.404391582799635e-06,
+      "loss": 0.6911,
+      "step": 140
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 6.86184812442818e-06,
+      "loss": 0.686,
+      "step": 150
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 7.319304666056725e-06,
+      "loss": 0.6844,
+      "step": 160
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 7.77676120768527e-06,
+      "loss": 0.687,
+      "step": 170
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 8.234217749313817e-06,
+      "loss": 0.6861,
+      "step": 180
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 8.691674290942362e-06,
+      "loss": 0.6797,
+      "step": 190
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.149130832570907e-06,
+      "loss": 0.6838,
+      "step": 200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.606587374199452e-06,
+      "loss": 0.6819,
+      "step": 210
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.0064043915827996e-05,
+      "loss": 0.6823,
+      "step": 220
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.0521500457456541e-05,
+      "loss": 0.6786,
+      "step": 230
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.0978956999085088e-05,
+      "loss": 0.6803,
+      "step": 240
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.1436413540713633e-05,
+      "loss": 0.6777,
+      "step": 250
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.1893870082342178e-05,
+      "loss": 0.675,
+      "step": 260
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.2351326623970723e-05,
+      "loss": 0.6775,
+      "step": 270
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.280878316559927e-05,
+      "loss": 0.6753,
+      "step": 280
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.3266239707227815e-05,
+      "loss": 0.672,
+      "step": 290
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.372369624885636e-05,
+      "loss": 0.6681,
+      "step": 300
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.4181152790484905e-05,
+      "loss": 0.6684,
+      "step": 310
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.463860933211345e-05,
+      "loss": 0.6629,
+      "step": 320
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.5096065873741997e-05,
+      "loss": 0.6662,
+      "step": 330
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.555352241537054e-05,
+      "loss": 0.663,
+      "step": 340
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6010978956999087e-05,
+      "loss": 0.6643,
+      "step": 350
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.6468435498627634e-05,
+      "loss": 0.6637,
+      "step": 360
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.6925892040256177e-05,
+      "loss": 0.6566,
+      "step": 370
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.7383348581884724e-05,
+      "loss": 0.65,
+      "step": 380
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.7840805123513268e-05,
+      "loss": 0.6485,
+      "step": 390
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8298261665141814e-05,
+      "loss": 0.6504,
+      "step": 400
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8755718206770358e-05,
+      "loss": 0.6424,
+      "step": 410
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9213174748398905e-05,
+      "loss": 0.6477,
+      "step": 420
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9670631290027448e-05,
+      "loss": 0.6381,
+      "step": 430
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.012808783165599e-05,
+      "loss": 0.6354,
+      "step": 440
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.0585544373284538e-05,
+      "loss": 0.6319,
+      "step": 450
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.1043000914913082e-05,
+      "loss": 0.6386,
+      "step": 460
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.150045745654163e-05,
+      "loss": 0.6318,
+      "step": 470
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.1957913998170175e-05,
+      "loss": 0.626,
+      "step": 480
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.241537053979872e-05,
+      "loss": 0.6232,
+      "step": 490
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.2872827081427266e-05,
+      "loss": 0.6239,
+      "step": 500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.333028362305581e-05,
+      "loss": 0.6057,
+      "step": 510
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.3787740164684356e-05,
+      "loss": 0.6112,
+      "step": 520
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.42451967063129e-05,
+      "loss": 0.6059,
+      "step": 530
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.4702653247941446e-05,
+      "loss": 0.5989,
+      "step": 540
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.5160109789569993e-05,
+      "loss": 0.5978,
+      "step": 550
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.561756633119854e-05,
+      "loss": 0.6082,
+      "step": 560
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.6075022872827083e-05,
+      "loss": 0.6,
+      "step": 570
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.653247941445563e-05,
+      "loss": 0.5772,
+      "step": 580
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.6989935956084173e-05,
+      "loss": 0.5737,
+      "step": 590
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.744739249771272e-05,
+      "loss": 0.5648,
+      "step": 600
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.7904849039341263e-05,
+      "loss": 0.5793,
+      "step": 610
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.836230558096981e-05,
+      "loss": 0.559,
+      "step": 620
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.8819762122598354e-05,
+      "loss": 0.5597,
+      "step": 630
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.92772186642269e-05,
+      "loss": 0.545,
+      "step": 640
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.9734675205855444e-05,
+      "loss": 0.5474,
+      "step": 650
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.0192131747483994e-05,
+      "loss": 0.5349,
+      "step": 660
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.064958828911254e-05,
+      "loss": 0.5188,
+      "step": 670
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.110704483074108e-05,
+      "loss": 0.5221,
+      "step": 680
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1564501372369624e-05,
+      "loss": 0.5109,
+      "step": 690
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.2021957913998174e-05,
+      "loss": 0.5262,
+      "step": 700
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.247941445562672e-05,
+      "loss": 0.5363,
+      "step": 710
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.293687099725527e-05,
+      "loss": 0.4986,
+      "step": 720
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.3394327538883805e-05,
+      "loss": 0.5339,
+      "step": 730
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.3851784080512355e-05,
+      "loss": 0.506,
+      "step": 740
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.43092406221409e-05,
+      "loss": 0.5039,
+      "step": 750
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.476669716376945e-05,
+      "loss": 0.5214,
+      "step": 760
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.522415370539799e-05,
+      "loss": 0.5038,
+      "step": 770
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.5681610247026535e-05,
+      "loss": 0.5083,
+      "step": 780
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.613906678865508e-05,
+      "loss": 0.4845,
+      "step": 790
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.659652333028363e-05,
+      "loss": 0.4859,
+      "step": 800
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.705397987191217e-05,
+      "loss": 0.4749,
+      "step": 810
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7511436413540716e-05,
+      "loss": 0.4769,
+      "step": 820
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.796889295516926e-05,
+      "loss": 0.4856,
+      "step": 830
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.842634949679781e-05,
+      "loss": 0.5029,
+      "step": 840
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.888380603842635e-05,
+      "loss": 0.437,
+      "step": 850
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.9341262580054896e-05,
+      "loss": 0.4531,
+      "step": 860
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.979871912168344e-05,
+      "loss": 0.4841,
+      "step": 870
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.025617566331198e-05,
+      "loss": 0.4457,
+      "step": 880
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.071363220494053e-05,
+      "loss": 0.5211,
+      "step": 890
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1171088746569077e-05,
+      "loss": 0.4364,
+      "step": 900
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.162854528819763e-05,
+      "loss": 0.4624,
+      "step": 910
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.2086001829826163e-05,
+      "loss": 0.4523,
+      "step": 920
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2543458371454714e-05,
+      "loss": 0.4297,
+      "step": 930
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.300091491308326e-05,
+      "loss": 0.4629,
+      "step": 940
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.345837145471181e-05,
+      "loss": 0.4264,
+      "step": 950
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.391582799634035e-05,
+      "loss": 0.4441,
+      "step": 960
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.4373284537968894e-05,
+      "loss": 0.4166,
+      "step": 970
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.483074107959744e-05,
+      "loss": 0.4265,
+      "step": 980
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.528819762122599e-05,
+      "loss": 0.4168,
+      "step": 990
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.574565416285453e-05,
+      "loss": 0.4514,
+      "step": 1000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.6203110704483074e-05,
+      "loss": 0.4106,
+      "step": 1010
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.666056724611162e-05,
+      "loss": 0.4104,
+      "step": 1020
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.711802378774017e-05,
+      "loss": 0.4546,
+      "step": 1030
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.757548032936871e-05,
+      "loss": 0.4559,
+      "step": 1040
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.8032936870997255e-05,
+      "loss": 0.399,
+      "step": 1050
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.84903934126258e-05,
+      "loss": 0.4097,
+      "step": 1060
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.894784995425435e-05,
+      "loss": 0.424,
+      "step": 1070
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.940530649588289e-05,
+      "loss": 0.3972,
+      "step": 1080
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.986276303751144e-05,
+      "loss": 0.441,
+      "step": 1090
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8458333333333333,
+      "eval_loss": 0.3588228225708008,
+      "eval_runtime": 74.7802,
+      "eval_samples_per_second": 401.176,
+      "eval_steps_per_second": 25.073,
+      "step": 1093
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.996442004676223e-05,
+      "loss": 0.4284,
+      "step": 1100
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.9913591542136836e-05,
+      "loss": 0.3938,
+      "step": 1110
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.986276303751144e-05,
+      "loss": 0.3838,
+      "step": 1120
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.981193453288605e-05,
+      "loss": 0.4133,
+      "step": 1130
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.976110602826065e-05,
+      "loss": 0.4262,
+      "step": 1140
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.9710277523635254e-05,
+      "loss": 0.4036,
+      "step": 1150
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.965944901900986e-05,
+      "loss": 0.3887,
+      "step": 1160
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.960862051438447e-05,
+      "loss": 0.4085,
+      "step": 1170
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.955779200975907e-05,
+      "loss": 0.4163,
+      "step": 1180
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.950696350513368e-05,
+      "loss": 0.4161,
+      "step": 1190
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.9456135000508286e-05,
+      "loss": 0.4192,
+      "step": 1200
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.940530649588289e-05,
+      "loss": 0.3895,
+      "step": 1210
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.93544779912575e-05,
+      "loss": 0.3681,
+      "step": 1220
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.9303649486632104e-05,
+      "loss": 0.3812,
+      "step": 1230
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.925282098200672e-05,
+      "loss": 0.4074,
+      "step": 1240
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.920199247738132e-05,
+      "loss": 0.3866,
+      "step": 1250
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.915116397275592e-05,
+      "loss": 0.3831,
+      "step": 1260
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.910033546813053e-05,
+      "loss": 0.3559,
+      "step": 1270
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.9049506963505136e-05,
+      "loss": 0.4156,
+      "step": 1280
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.899867845887974e-05,
+      "loss": 0.3929,
+      "step": 1290
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.894784995425435e-05,
+      "loss": 0.4173,
+      "step": 1300
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.8897021449628955e-05,
+      "loss": 0.4122,
+      "step": 1310
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.884619294500356e-05,
+      "loss": 0.4078,
+      "step": 1320
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.879536444037817e-05,
+      "loss": 0.348,
+      "step": 1330
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.8744535935752774e-05,
+      "loss": 0.3975,
+      "step": 1340
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.869370743112738e-05,
+      "loss": 0.3858,
+      "step": 1350
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.8642878926501986e-05,
+      "loss": 0.3828,
+      "step": 1360
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.859205042187659e-05,
+      "loss": 0.3595,
+      "step": 1370
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.854122191725119e-05,
+      "loss": 0.37,
+      "step": 1380
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.84903934126258e-05,
+      "loss": 0.3779,
+      "step": 1390
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.8439564908000405e-05,
+      "loss": 0.383,
+      "step": 1400
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.838873640337501e-05,
+      "loss": 0.3889,
+      "step": 1410
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.8337907898749624e-05,
+      "loss": 0.3617,
+      "step": 1420
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.828707939412423e-05,
+      "loss": 0.3722,
+      "step": 1430
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.8236250889498837e-05,
+      "loss": 0.3462,
+      "step": 1440
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.818542238487344e-05,
+      "loss": 0.4152,
+      "step": 1450
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.813459388024805e-05,
+      "loss": 0.3959,
+      "step": 1460
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.8083765375622655e-05,
+      "loss": 0.3515,
+      "step": 1470
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.8032936870997255e-05,
+      "loss": 0.3762,
+      "step": 1480
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.798210836637186e-05,
+      "loss": 0.3722,
+      "step": 1490
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.793127986174647e-05,
+      "loss": 0.3592,
+      "step": 1500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.7880451357121074e-05,
+      "loss": 0.3577,
+      "step": 1510
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.782962285249568e-05,
+      "loss": 0.3893,
+      "step": 1520
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.7778794347870286e-05,
+      "loss": 0.3896,
+      "step": 1530
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.772796584324489e-05,
+      "loss": 0.3638,
+      "step": 1540
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.76771373386195e-05,
+      "loss": 0.3395,
+      "step": 1550
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.7626308833994105e-05,
+      "loss": 0.3683,
+      "step": 1560
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.757548032936871e-05,
+      "loss": 0.3677,
+      "step": 1570
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.752465182474332e-05,
+      "loss": 0.3731,
+      "step": 1580
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.7473823320117924e-05,
+      "loss": 0.3564,
+      "step": 1590
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.742299481549253e-05,
+      "loss": 0.3535,
+      "step": 1600
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.737216631086714e-05,
+      "loss": 0.3741,
+      "step": 1610
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.732133780624174e-05,
+      "loss": 0.3634,
+      "step": 1620
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.727050930161635e-05,
+      "loss": 0.3786,
+      "step": 1630
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.7219680796990955e-05,
+      "loss": 0.3418,
+      "step": 1640
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.716885229236556e-05,
+      "loss": 0.3537,
+      "step": 1650
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.711802378774017e-05,
+      "loss": 0.3456,
+      "step": 1660
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.7067195283114774e-05,
+      "loss": 0.3948,
+      "step": 1670
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.701636677848938e-05,
+      "loss": 0.3506,
+      "step": 1680
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.696553827386399e-05,
+      "loss": 0.3698,
+      "step": 1690
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.691470976923859e-05,
+      "loss": 0.3706,
+      "step": 1700
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.68638812646132e-05,
+      "loss": 0.3522,
+      "step": 1710
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.68130527599878e-05,
+      "loss": 0.4005,
+      "step": 1720
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.6762224255362405e-05,
+      "loss": 0.3412,
+      "step": 1730
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.671139575073701e-05,
+      "loss": 0.3277,
+      "step": 1740
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.666056724611162e-05,
+      "loss": 0.3756,
+      "step": 1750
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.6609738741486224e-05,
+      "loss": 0.408,
+      "step": 1760
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.655891023686084e-05,
+      "loss": 0.3762,
+      "step": 1770
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.6508081732235444e-05,
+      "loss": 0.3269,
+      "step": 1780
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.645725322761005e-05,
+      "loss": 0.3394,
+      "step": 1790
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.6406424722984656e-05,
+      "loss": 0.3724,
+      "step": 1800
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.635559621835926e-05,
+      "loss": 0.3655,
+      "step": 1810
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.630476771373387e-05,
+      "loss": 0.3734,
+      "step": 1820
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.625393920910847e-05,
+      "loss": 0.3454,
+      "step": 1830
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.6203110704483074e-05,
+      "loss": 0.3474,
+      "step": 1840
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.615228219985768e-05,
+      "loss": 0.3598,
+      "step": 1850
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.610145369523229e-05,
+      "loss": 0.3624,
+      "step": 1860
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.605062519060689e-05,
+      "loss": 0.3321,
+      "step": 1870
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.59997966859815e-05,
+      "loss": 0.3124,
+      "step": 1880
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.5948968181356106e-05,
+      "loss": 0.3545,
+      "step": 1890
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.589813967673071e-05,
+      "loss": 0.4075,
+      "step": 1900
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.584731117210532e-05,
+      "loss": 0.3382,
+      "step": 1910
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.5796482667479925e-05,
+      "loss": 0.3668,
+      "step": 1920
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.574565416285453e-05,
+      "loss": 0.3247,
+      "step": 1930
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.569482565822914e-05,
+      "loss": 0.3424,
+      "step": 1940
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.5643997153603744e-05,
+      "loss": 0.3095,
+      "step": 1950
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.559316864897835e-05,
+      "loss": 0.3061,
+      "step": 1960
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.5542340144352956e-05,
+      "loss": 0.3598,
+      "step": 1970
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.549151163972756e-05,
+      "loss": 0.3654,
+      "step": 1980
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.544068313510217e-05,
+      "loss": 0.3404,
+      "step": 1990
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.5389854630476775e-05,
+      "loss": 0.3478,
+      "step": 2000
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.533902612585138e-05,
+      "loss": 0.3552,
+      "step": 2010
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.528819762122599e-05,
+      "loss": 0.339,
+      "step": 2020
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.5237369116600594e-05,
+      "loss": 0.3643,
+      "step": 2030
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.51865406119752e-05,
+      "loss": 0.3191,
+      "step": 2040
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.5135712107349806e-05,
+      "loss": 0.3477,
+      "step": 2050
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.5084883602724406e-05,
+      "loss": 0.3382,
+      "step": 2060
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.503405509809901e-05,
+      "loss": 0.3242,
+      "step": 2070
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.498322659347362e-05,
+      "loss": 0.3571,
+      "step": 2080
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.4932398088848225e-05,
+      "loss": 0.3587,
+      "step": 2090
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.488156958422283e-05,
+      "loss": 0.3218,
+      "step": 2100
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.483074107959744e-05,
+      "loss": 0.3351,
+      "step": 2110
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.4779912574972044e-05,
+      "loss": 0.3303,
+      "step": 2120
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.472908407034666e-05,
+      "loss": 0.333,
+      "step": 2130
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.467825556572126e-05,
+      "loss": 0.3287,
+      "step": 2140
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.462742706109587e-05,
+      "loss": 0.359,
+      "step": 2150
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.4576598556470476e-05,
+      "loss": 0.3351,
+      "step": 2160
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.4525770051845075e-05,
+      "loss": 0.3382,
+      "step": 2170
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.447494154721968e-05,
+      "loss": 0.2977,
+      "step": 2180
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8806333333333334,
+      "eval_loss": 0.2797642648220062,
+      "eval_runtime": 73.6465,
+      "eval_samples_per_second": 407.351,
+      "eval_steps_per_second": 25.459,
+      "step": 2187
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.442411304259429e-05,
+      "loss": 0.3467,
+      "step": 2190
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.4373284537968894e-05,
+      "loss": 0.3314,
+      "step": 2200
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.43224560333435e-05,
+      "loss": 0.348,
+      "step": 2210
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.4271627528718107e-05,
+      "loss": 0.3237,
+      "step": 2220
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.422079902409271e-05,
+      "loss": 0.2997,
+      "step": 2230
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.416997051946732e-05,
+      "loss": 0.3352,
+      "step": 2240
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.4119142014841925e-05,
+      "loss": 0.3091,
+      "step": 2250
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.406831351021653e-05,
+      "loss": 0.3481,
+      "step": 2260
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.401748500559114e-05,
+      "loss": 0.3393,
+      "step": 2270
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.3966656500965744e-05,
+      "loss": 0.3444,
+      "step": 2280
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.391582799634035e-05,
+      "loss": 0.3401,
+      "step": 2290
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.386499949171495e-05,
+      "loss": 0.3266,
+      "step": 2300
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.381417098708956e-05,
+      "loss": 0.3452,
+      "step": 2310
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.376334248246417e-05,
+      "loss": 0.2973,
+      "step": 2320
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.3712513977838776e-05,
+      "loss": 0.3373,
+      "step": 2330
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.366168547321338e-05,
+      "loss": 0.3282,
+      "step": 2340
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.361085696858799e-05,
+      "loss": 0.3439,
+      "step": 2350
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.3560028463962595e-05,
+      "loss": 0.312,
+      "step": 2360
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.35091999593372e-05,
+      "loss": 0.3598,
+      "step": 2370
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.345837145471181e-05,
+      "loss": 0.3163,
+      "step": 2380
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.3407542950086413e-05,
+      "loss": 0.3042,
+      "step": 2390
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.335671444546101e-05,
+      "loss": 0.3695,
+      "step": 2400
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.330588594083562e-05,
+      "loss": 0.3025,
+      "step": 2410
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.3255057436210226e-05,
+      "loss": 0.3365,
+      "step": 2420
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.320422893158483e-05,
+      "loss": 0.3135,
+      "step": 2430
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.315340042695944e-05,
+      "loss": 0.3314,
+      "step": 2440
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.3102571922334044e-05,
+      "loss": 0.3387,
+      "step": 2450
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.305174341770865e-05,
+      "loss": 0.3128,
+      "step": 2460
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.300091491308326e-05,
+      "loss": 0.3186,
+      "step": 2470
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.295008640845786e-05,
+      "loss": 0.3166,
+      "step": 2480
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.2899257903832476e-05,
+      "loss": 0.3201,
+      "step": 2490
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.284842939920708e-05,
+      "loss": 0.2944,
+      "step": 2500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.279760089458168e-05,
+      "loss": 0.2983,
+      "step": 2510
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.274677238995629e-05,
+      "loss": 0.3445,
+      "step": 2520
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.2695943885330895e-05,
+      "loss": 0.3487,
+      "step": 2530
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.26451153807055e-05,
+      "loss": 0.3101,
+      "step": 2540
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.259428687608011e-05,
+      "loss": 0.3435,
+      "step": 2550
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.2543458371454714e-05,
+      "loss": 0.2977,
+      "step": 2560
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.249262986682932e-05,
+      "loss": 0.3403,
+      "step": 2570
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.2441801362203926e-05,
+      "loss": 0.3271,
+      "step": 2580
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.239097285757853e-05,
+      "loss": 0.3044,
+      "step": 2590
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.234014435295314e-05,
+      "loss": 0.3404,
+      "step": 2600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.2289315848327745e-05,
+      "loss": 0.3133,
+      "step": 2610
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.223848734370235e-05,
+      "loss": 0.3215,
+      "step": 2620
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.218765883907696e-05,
+      "loss": 0.3233,
+      "step": 2630
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.213683033445156e-05,
+      "loss": 0.3818,
+      "step": 2640
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.2086001829826163e-05,
+      "loss": 0.2774,
+      "step": 2650
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.2035173325200776e-05,
+      "loss": 0.2957,
+      "step": 2660
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.198434482057538e-05,
+      "loss": 0.3131,
+      "step": 2670
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.193351631594999e-05,
+      "loss": 0.3616,
+      "step": 2680
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.1882687811324595e-05,
+      "loss": 0.3189,
+      "step": 2690
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.18318593066992e-05,
+      "loss": 0.2773,
+      "step": 2700
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 4.178103080207381e-05,
+      "loss": 0.338,
+      "step": 2710
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 4.1730202297448414e-05,
+      "loss": 0.3245,
+      "step": 2720
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 4.167937379282302e-05,
+      "loss": 0.3093,
+      "step": 2730
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.162854528819763e-05,
+      "loss": 0.3354,
+      "step": 2740
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1577716783572226e-05,
+      "loss": 0.3541,
+      "step": 2750
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.152688827894683e-05,
+      "loss": 0.3295,
+      "step": 2760
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 4.147605977432144e-05,
+      "loss": 0.3322,
+      "step": 2770
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.1425231269696045e-05,
+      "loss": 0.3532,
+      "step": 2780
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.137440276507065e-05,
+      "loss": 0.28,
+      "step": 2790
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.132357426044526e-05,
+      "loss": 0.3287,
+      "step": 2800
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 4.1272745755819864e-05,
+      "loss": 0.2997,
+      "step": 2810
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.122191725119447e-05,
+      "loss": 0.3167,
+      "step": 2820
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.1171088746569077e-05,
+      "loss": 0.362,
+      "step": 2830
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.112026024194369e-05,
+      "loss": 0.3173,
+      "step": 2840
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.106943173731829e-05,
+      "loss": 0.3161,
+      "step": 2850
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.1018603232692895e-05,
+      "loss": 0.3226,
+      "step": 2860
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.09677747280675e-05,
+      "loss": 0.3277,
+      "step": 2870
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.091694622344211e-05,
+      "loss": 0.3172,
+      "step": 2880
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.0866117718816714e-05,
+      "loss": 0.3385,
+      "step": 2890
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.081528921419132e-05,
+      "loss": 0.3616,
+      "step": 2900
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 4.076446070956593e-05,
+      "loss": 0.3073,
+      "step": 2910
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.071363220494053e-05,
+      "loss": 0.2994,
+      "step": 2920
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 4.066280370031514e-05,
+      "loss": 0.3233,
+      "step": 2930
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.0611975195689746e-05,
+      "loss": 0.3405,
+      "step": 2940
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 4.056114669106435e-05,
+      "loss": 0.3447,
+      "step": 2950
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.051031818643896e-05,
+      "loss": 0.2891,
+      "step": 2960
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.0459489681813565e-05,
+      "loss": 0.2644,
+      "step": 2970
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 4.0408661177188164e-05,
+      "loss": 0.3364,
+      "step": 2980
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.035783267256277e-05,
+      "loss": 0.297,
+      "step": 2990
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 4.030700416793738e-05,
+      "loss": 0.2779,
+      "step": 3000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.025617566331198e-05,
+      "loss": 0.3317,
+      "step": 3010
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 4.0205347158686596e-05,
+      "loss": 0.3772,
+      "step": 3020
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.01545186540612e-05,
+      "loss": 0.3045,
+      "step": 3030
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.010369014943581e-05,
+      "loss": 0.3151,
+      "step": 3040
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.0052861644810415e-05,
+      "loss": 0.291,
+      "step": 3050
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.000203314018502e-05,
+      "loss": 0.2781,
+      "step": 3060
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 3.995120463555963e-05,
+      "loss": 0.3246,
+      "step": 3070
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3.9900376130934234e-05,
+      "loss": 0.3133,
+      "step": 3080
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.984954762630883e-05,
+      "loss": 0.3019,
+      "step": 3090
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.979871912168344e-05,
+      "loss": 0.3227,
+      "step": 3100
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.9747890617058046e-05,
+      "loss": 0.3117,
+      "step": 3110
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.969706211243265e-05,
+      "loss": 0.3344,
+      "step": 3120
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 3.964623360780726e-05,
+      "loss": 0.3054,
+      "step": 3130
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 3.9595405103181865e-05,
+      "loss": 0.3451,
+      "step": 3140
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 3.954457659855647e-05,
+      "loss": 0.2879,
+      "step": 3150
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.949374809393108e-05,
+      "loss": 0.2933,
+      "step": 3160
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.9442919589305684e-05,
+      "loss": 0.3139,
+      "step": 3170
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 3.939209108468029e-05,
+      "loss": 0.2863,
+      "step": 3180
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 3.9341262580054896e-05,
+      "loss": 0.3213,
+      "step": 3190
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.92904340754295e-05,
+      "loss": 0.3213,
+      "step": 3200
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.923960557080411e-05,
+      "loss": 0.3252,
+      "step": 3210
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 3.9188777066178715e-05,
+      "loss": 0.3112,
+      "step": 3220
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.913794856155332e-05,
+      "loss": 0.336,
+      "step": 3230
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 3.908712005692793e-05,
+      "loss": 0.2882,
+      "step": 3240
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.9036291552302534e-05,
+      "loss": 0.3221,
+      "step": 3250
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.898546304767714e-05,
+      "loss": 0.3202,
+      "step": 3260
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.8934634543051746e-05,
+      "loss": 0.3281,
+      "step": 3270
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.888380603842635e-05,
+      "loss": 0.3006,
+      "step": 3280
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9221666666666667,
+      "eval_loss": 0.19573450088500977,
+      "eval_runtime": 73.3697,
+      "eval_samples_per_second": 408.888,
+      "eval_steps_per_second": 25.556,
+      "step": 3281
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.883297753380096e-05,
+      "loss": 0.2654,
+      "step": 3290
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 3.8782149029175565e-05,
+      "loss": 0.3032,
+      "step": 3300
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.873132052455017e-05,
+      "loss": 0.2797,
+      "step": 3310
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.868049201992477e-05,
+      "loss": 0.2948,
+      "step": 3320
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.862966351529938e-05,
+      "loss": 0.297,
+      "step": 3330
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.8578835010673984e-05,
+      "loss": 0.3011,
+      "step": 3340
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 3.852800650604859e-05,
+      "loss": 0.3059,
+      "step": 3350
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.8477178001423196e-05,
+      "loss": 0.2889,
+      "step": 3360
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.842634949679781e-05,
+      "loss": 0.3551,
+      "step": 3370
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.8375520992172416e-05,
+      "loss": 0.3206,
+      "step": 3380
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.832469248754702e-05,
+      "loss": 0.3318,
+      "step": 3390
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.827386398292163e-05,
+      "loss": 0.3342,
+      "step": 3400
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.8223035478296234e-05,
+      "loss": 0.2808,
+      "step": 3410
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.817220697367084e-05,
+      "loss": 0.2679,
+      "step": 3420
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.812137846904544e-05,
+      "loss": 0.3173,
+      "step": 3430
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.8070549964420047e-05,
+      "loss": 0.3255,
+      "step": 3440
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.801972145979465e-05,
+      "loss": 0.2776,
+      "step": 3450
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.796889295516926e-05,
+      "loss": 0.3016,
+      "step": 3460
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.7918064450543865e-05,
+      "loss": 0.2932,
+      "step": 3470
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.786723594591847e-05,
+      "loss": 0.3169,
+      "step": 3480
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.781640744129308e-05,
+      "loss": 0.3319,
+      "step": 3490
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.7765578936667684e-05,
+      "loss": 0.2785,
+      "step": 3500
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.771475043204229e-05,
+      "loss": 0.3365,
+      "step": 3510
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 3.76639219274169e-05,
+      "loss": 0.2877,
+      "step": 3520
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.76130934227915e-05,
+      "loss": 0.29,
+      "step": 3530
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.756226491816611e-05,
+      "loss": 0.2735,
+      "step": 3540
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.7511436413540716e-05,
+      "loss": 0.2795,
+      "step": 3550
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.746060790891532e-05,
+      "loss": 0.2637,
+      "step": 3560
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.740977940428993e-05,
+      "loss": 0.2861,
+      "step": 3570
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.7358950899664535e-05,
+      "loss": 0.2838,
+      "step": 3580
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.730812239503914e-05,
+      "loss": 0.3116,
+      "step": 3590
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.725729389041375e-05,
+      "loss": 0.3189,
+      "step": 3600
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.7206465385788353e-05,
+      "loss": 0.3303,
+      "step": 3610
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.715563688116296e-05,
+      "loss": 0.3001,
+      "step": 3620
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.7104808376537566e-05,
+      "loss": 0.3311,
+      "step": 3630
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.705397987191217e-05,
+      "loss": 0.2484,
+      "step": 3640
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.700315136728678e-05,
+      "loss": 0.3047,
+      "step": 3650
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.6952322862661385e-05,
+      "loss": 0.273,
+      "step": 3660
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.6901494358035984e-05,
+      "loss": 0.3088,
+      "step": 3670
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.685066585341059e-05,
+      "loss": 0.2751,
+      "step": 3680
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.67998373487852e-05,
+      "loss": 0.3388,
+      "step": 3690
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.67490088441598e-05,
+      "loss": 0.2858,
+      "step": 3700
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.669818033953441e-05,
+      "loss": 0.3009,
+      "step": 3710
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.6647351834909016e-05,
+      "loss": 0.2915,
+      "step": 3720
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.659652333028363e-05,
+      "loss": 0.2834,
+      "step": 3730
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.6545694825658235e-05,
+      "loss": 0.3081,
+      "step": 3740
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.649486632103284e-05,
+      "loss": 0.3201,
+      "step": 3750
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.644403781640745e-05,
+      "loss": 0.3235,
+      "step": 3760
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 3.639320931178205e-05,
+      "loss": 0.2966,
+      "step": 3770
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.6342380807156654e-05,
+      "loss": 0.3387,
+      "step": 3780
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.629155230253126e-05,
+      "loss": 0.3157,
+      "step": 3790
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.6240723797905866e-05,
+      "loss": 0.3029,
+      "step": 3800
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.618989529328047e-05,
+      "loss": 0.3228,
+      "step": 3810
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 3.613906678865508e-05,
+      "loss": 0.3458,
+      "step": 3820
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.6088238284029685e-05,
+      "loss": 0.2915,
+      "step": 3830
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.603740977940429e-05,
+      "loss": 0.3271,
+      "step": 3840
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.59865812747789e-05,
+      "loss": 0.3215,
+      "step": 3850
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.5935752770153504e-05,
+      "loss": 0.2742,
+      "step": 3860
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 3.588492426552811e-05,
+      "loss": 0.32,
+      "step": 3870
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.5834095760902716e-05,
+      "loss": 0.2787,
+      "step": 3880
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 3.578326725627732e-05,
+      "loss": 0.3046,
+      "step": 3890
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.573243875165192e-05,
+      "loss": 0.3581,
+      "step": 3900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.5681610247026535e-05,
+      "loss": 0.2798,
+      "step": 3910
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.563078174240114e-05,
+      "loss": 0.309,
+      "step": 3920
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.557995323777575e-05,
+      "loss": 0.3269,
+      "step": 3930
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.5529124733150354e-05,
+      "loss": 0.2835,
+      "step": 3940
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.547829622852496e-05,
+      "loss": 0.2783,
+      "step": 3950
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.542746772389957e-05,
+      "loss": 0.2858,
+      "step": 3960
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.537663921927417e-05,
+      "loss": 0.3268,
+      "step": 3970
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.532581071464878e-05,
+      "loss": 0.3005,
+      "step": 3980
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.5274982210023386e-05,
+      "loss": 0.2914,
+      "step": 3990
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.522415370539799e-05,
+      "loss": 0.2898,
+      "step": 4000
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.517332520077259e-05,
+      "loss": 0.2892,
+      "step": 4010
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.51224966961472e-05,
+      "loss": 0.3086,
+      "step": 4020
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.5071668191521804e-05,
+      "loss": 0.3182,
+      "step": 4030
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.502083968689641e-05,
+      "loss": 0.3217,
+      "step": 4040
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.4970011182271017e-05,
+      "loss": 0.2846,
+      "step": 4050
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.491918267764562e-05,
+      "loss": 0.3009,
+      "step": 4060
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.486835417302023e-05,
+      "loss": 0.2651,
+      "step": 4070
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.4817525668394835e-05,
+      "loss": 0.3328,
+      "step": 4080
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.476669716376945e-05,
+      "loss": 0.2846,
+      "step": 4090
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4715868659144055e-05,
+      "loss": 0.2708,
+      "step": 4100
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.4665040154518654e-05,
+      "loss": 0.2744,
+      "step": 4110
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.461421164989326e-05,
+      "loss": 0.2651,
+      "step": 4120
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.456338314526787e-05,
+      "loss": 0.3104,
+      "step": 4130
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.451255464064247e-05,
+      "loss": 0.3142,
+      "step": 4140
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.446172613601708e-05,
+      "loss": 0.3287,
+      "step": 4150
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.4410897631391686e-05,
+      "loss": 0.2851,
+      "step": 4160
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.436006912676629e-05,
+      "loss": 0.2862,
+      "step": 4170
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.43092406221409e-05,
+      "loss": 0.3324,
+      "step": 4180
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.4258412117515505e-05,
+      "loss": 0.3068,
+      "step": 4190
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.420758361289011e-05,
+      "loss": 0.2988,
+      "step": 4200
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.415675510826472e-05,
+      "loss": 0.2955,
+      "step": 4210
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.4105926603639323e-05,
+      "loss": 0.2816,
+      "step": 4220
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.405509809901393e-05,
+      "loss": 0.3027,
+      "step": 4230
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.400426959438853e-05,
+      "loss": 0.3071,
+      "step": 4240
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.3953441089763136e-05,
+      "loss": 0.3399,
+      "step": 4250
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.390261258513775e-05,
+      "loss": 0.2729,
+      "step": 4260
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.3851784080512355e-05,
+      "loss": 0.2492,
+      "step": 4270
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.380095557588696e-05,
+      "loss": 0.2826,
+      "step": 4280
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 3.375012707126157e-05,
+      "loss": 0.3025,
+      "step": 4290
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.3699298566636174e-05,
+      "loss": 0.2673,
+      "step": 4300
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.364847006201078e-05,
+      "loss": 0.3235,
+      "step": 4310
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.3597641557385386e-05,
+      "loss": 0.2737,
+      "step": 4320
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.354681305275999e-05,
+      "loss": 0.2944,
+      "step": 4330
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.34959845481346e-05,
+      "loss": 0.3253,
+      "step": 4340
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.34451560435092e-05,
+      "loss": 0.2629,
+      "step": 4350
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 3.3394327538883805e-05,
+      "loss": 0.2892,
+      "step": 4360
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.334349903425841e-05,
+      "loss": 0.2916,
+      "step": 4370
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9323,
+      "eval_loss": 0.17996086180210114,
+      "eval_runtime": 74.1638,
+      "eval_samples_per_second": 404.51,
+      "eval_steps_per_second": 25.282,
+      "step": 4375
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.329267052963302e-05,
+      "loss": 0.3261,
+      "step": 4380
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 3.3241842025007624e-05,
+      "loss": 0.2857,
+      "step": 4390
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 3.319101352038223e-05,
+      "loss": 0.2695,
+      "step": 4400
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 3.3140185015756836e-05,
+      "loss": 0.3276,
+      "step": 4410
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 3.308935651113144e-05,
+      "loss": 0.2874,
+      "step": 4420
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 3.303852800650605e-05,
+      "loss": 0.3304,
+      "step": 4430
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 3.298769950188066e-05,
+      "loss": 0.3024,
+      "step": 4440
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.293687099725527e-05,
+      "loss": 0.3395,
+      "step": 4450
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 3.288604249262987e-05,
+      "loss": 0.3108,
+      "step": 4460
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 3.2835213988004474e-05,
+      "loss": 0.2874,
+      "step": 4470
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 3.278438548337908e-05,
+      "loss": 0.3035,
+      "step": 4480
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 3.2733556978753686e-05,
+      "loss": 0.2544,
+      "step": 4490
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 3.268272847412829e-05,
+      "loss": 0.2819,
+      "step": 4500
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 3.26318999695029e-05,
+      "loss": 0.351,
+      "step": 4510
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 3.2581071464877505e-05,
+      "loss": 0.2544,
+      "step": 4520
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 3.253024296025211e-05,
+      "loss": 0.2666,
+      "step": 4530
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 3.247941445562672e-05,
+      "loss": 0.2849,
+      "step": 4540
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 3.2428585951001324e-05,
+      "loss": 0.288,
+      "step": 4550
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 3.237775744637593e-05,
+      "loss": 0.2764,
+      "step": 4560
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 3.232692894175054e-05,
+      "loss": 0.2883,
+      "step": 4570
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 3.227610043712514e-05,
+      "loss": 0.2849,
+      "step": 4580
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 3.222527193249974e-05,
+      "loss": 0.2797,
+      "step": 4590
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 3.217444342787435e-05,
+      "loss": 0.2512,
+      "step": 4600
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 3.2123614923248955e-05,
+      "loss": 0.2808,
+      "step": 4610
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 3.207278641862357e-05,
+      "loss": 0.2484,
+      "step": 4620
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 3.2021957913998174e-05,
+      "loss": 0.2465,
+      "step": 4630
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 3.197112940937278e-05,
+      "loss": 0.2659,
+      "step": 4640
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 3.192030090474739e-05,
+      "loss": 0.2737,
+      "step": 4650
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 3.186947240012199e-05,
+      "loss": 0.3075,
+      "step": 4660
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 3.18186438954966e-05,
+      "loss": 0.275,
+      "step": 4670
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 3.1767815390871206e-05,
+      "loss": 0.3339,
+      "step": 4680
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 3.1716986886245805e-05,
+      "loss": 0.27,
+      "step": 4690
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 3.166615838162041e-05,
+      "loss": 0.2873,
+      "step": 4700
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 3.161532987699502e-05,
+      "loss": 0.2709,
+      "step": 4710
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 3.1564501372369624e-05,
+      "loss": 0.3222,
+      "step": 4720
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 3.151367286774423e-05,
+      "loss": 0.2689,
+      "step": 4730
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 3.146284436311884e-05,
+      "loss": 0.2517,
+      "step": 4740
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 3.141201585849344e-05,
+      "loss": 0.2994,
+      "step": 4750
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 3.136118735386805e-05,
+      "loss": 0.2797,
+      "step": 4760
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 3.1310358849242656e-05,
+      "loss": 0.2836,
+      "step": 4770
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 3.125953034461726e-05,
+      "loss": 0.2909,
+      "step": 4780
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 3.120870183999187e-05,
+      "loss": 0.2415,
+      "step": 4790
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 3.1157873335366475e-05,
+      "loss": 0.2988,
+      "step": 4800
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 3.110704483074108e-05,
+      "loss": 0.2673,
+      "step": 4810
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 3.105621632611569e-05,
+      "loss": 0.2722,
+      "step": 4820
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 3.1005387821490293e-05,
+      "loss": 0.2685,
+      "step": 4830
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.09545593168649e-05,
+      "loss": 0.26,
+      "step": 4840
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.0903730812239506e-05,
+      "loss": 0.2852,
+      "step": 4850
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 3.085290230761411e-05,
+      "loss": 0.2619,
+      "step": 4860
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 3.080207380298872e-05,
+      "loss": 0.2737,
+      "step": 4870
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 3.0751245298363325e-05,
+      "loss": 0.2884,
+      "step": 4880
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 3.070041679373793e-05,
+      "loss": 0.3113,
+      "step": 4890
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 3.064958828911254e-05,
+      "loss": 0.2613,
+      "step": 4900
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 3.0598759784487144e-05,
+      "loss": 0.3106,
+      "step": 4910
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 3.054793127986175e-05,
+      "loss": 0.3016,
+      "step": 4920
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 3.0497102775236353e-05,
+      "loss": 0.2846,
+      "step": 4930
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 3.044627427061096e-05,
+      "loss": 0.2735,
+      "step": 4940
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 3.0395445765985565e-05,
+      "loss": 0.2401,
+      "step": 4950
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 3.034461726136017e-05,
+      "loss": 0.2792,
+      "step": 4960
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 3.0293788756734775e-05,
+      "loss": 0.2797,
+      "step": 4970
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 3.0242960252109388e-05,
+      "loss": 0.2652,
+      "step": 4980
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 3.0192131747483994e-05,
+      "loss": 0.2711,
+      "step": 4990
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 3.0141303242858597e-05,
+      "loss": 0.2602,
+      "step": 5000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.0090474738233203e-05,
+      "loss": 0.2992,
+      "step": 5010
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.003964623360781e-05,
+      "loss": 0.2547,
+      "step": 5020
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 2.9988817728982416e-05,
+      "loss": 0.2541,
+      "step": 5030
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 2.9937989224357022e-05,
+      "loss": 0.3136,
+      "step": 5040
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 2.988716071973163e-05,
+      "loss": 0.2734,
+      "step": 5050
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 2.9836332215106235e-05,
+      "loss": 0.258,
+      "step": 5060
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.9785503710480838e-05,
+      "loss": 0.244,
+      "step": 5070
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.9734675205855444e-05,
+      "loss": 0.241,
+      "step": 5080
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.968384670123005e-05,
+      "loss": 0.2653,
+      "step": 5090
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.9633018196604656e-05,
+      "loss": 0.2493,
+      "step": 5100
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.9582189691979263e-05,
+      "loss": 0.3328,
+      "step": 5110
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.953136118735387e-05,
+      "loss": 0.3044,
+      "step": 5120
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.9480532682728472e-05,
+      "loss": 0.3145,
+      "step": 5130
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.9429704178103078e-05,
+      "loss": 0.2888,
+      "step": 5140
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 2.937887567347769e-05,
+      "loss": 0.2914,
+      "step": 5150
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 2.9328047168852298e-05,
+      "loss": 0.2831,
+      "step": 5160
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 2.92772186642269e-05,
+      "loss": 0.2905,
+      "step": 5170
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 2.9226390159601507e-05,
+      "loss": 0.322,
+      "step": 5180
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.9175561654976113e-05,
+      "loss": 0.3045,
+      "step": 5190
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.912473315035072e-05,
+      "loss": 0.3134,
+      "step": 5200
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 2.9073904645725326e-05,
+      "loss": 0.2773,
+      "step": 5210
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 2.9023076141099932e-05,
+      "loss": 0.2738,
+      "step": 5220
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 2.8972247636474538e-05,
+      "loss": 0.2691,
+      "step": 5230
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 2.892141913184914e-05,
+      "loss": 0.2513,
+      "step": 5240
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 2.8870590627223747e-05,
+      "loss": 0.3047,
+      "step": 5250
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 2.8819762122598354e-05,
+      "loss": 0.245,
+      "step": 5260
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.876893361797296e-05,
+      "loss": 0.3325,
+      "step": 5270
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.8718105113347566e-05,
+      "loss": 0.2468,
+      "step": 5280
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.8667276608722172e-05,
+      "loss": 0.3309,
+      "step": 5290
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.8616448104096775e-05,
+      "loss": 0.286,
+      "step": 5300
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.856561959947138e-05,
+      "loss": 0.2509,
+      "step": 5310
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.8514791094845988e-05,
+      "loss": 0.2629,
+      "step": 5320
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.84639625902206e-05,
+      "loss": 0.2897,
+      "step": 5330
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.8413134085595204e-05,
+      "loss": 0.2778,
+      "step": 5340
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.836230558096981e-05,
+      "loss": 0.2987,
+      "step": 5350
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.8311477076344416e-05,
+      "loss": 0.2466,
+      "step": 5360
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.8260648571719023e-05,
+      "loss": 0.2894,
+      "step": 5370
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.820982006709363e-05,
+      "loss": 0.2495,
+      "step": 5380
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.8158991562468235e-05,
+      "loss": 0.28,
+      "step": 5390
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.810816305784284e-05,
+      "loss": 0.2555,
+      "step": 5400
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.8057334553217445e-05,
+      "loss": 0.3215,
+      "step": 5410
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.800650604859205e-05,
+      "loss": 0.2758,
+      "step": 5420
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.7955677543966657e-05,
+      "loss": 0.2745,
+      "step": 5430
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.7904849039341263e-05,
+      "loss": 0.2452,
+      "step": 5440
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.785402053471587e-05,
+      "loss": 0.2701,
+      "step": 5450
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.7803192030090476e-05,
+      "loss": 0.2835,
+      "step": 5460
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9306666666666666,
+      "eval_loss": 0.17844846844673157,
+      "eval_runtime": 74.7619,
+      "eval_samples_per_second": 401.274,
+      "eval_steps_per_second": 25.08,
+      "step": 5468
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.775236352546508e-05,
+      "loss": 0.2312,
+      "step": 5470
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 2.7701535020839685e-05,
+      "loss": 0.2698,
+      "step": 5480
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 2.765070651621429e-05,
+      "loss": 0.2697,
+      "step": 5490
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.7599878011588898e-05,
+      "loss": 0.2719,
+      "step": 5500
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 2.7549049506963507e-05,
+      "loss": 0.2659,
+      "step": 5510
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 2.7498221002338114e-05,
+      "loss": 0.2811,
+      "step": 5520
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 2.744739249771272e-05,
+      "loss": 0.2608,
+      "step": 5530
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.7396563993087326e-05,
+      "loss": 0.2531,
+      "step": 5540
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.7345735488461933e-05,
+      "loss": 0.2683,
+      "step": 5550
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 2.729490698383654e-05,
+      "loss": 0.303,
+      "step": 5560
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 2.7244078479211145e-05,
+      "loss": 0.2795,
+      "step": 5570
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.7193249974585748e-05,
+      "loss": 0.272,
+      "step": 5580
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 2.7142421469960354e-05,
+      "loss": 0.2534,
+      "step": 5590
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 2.709159296533496e-05,
+      "loss": 0.2941,
+      "step": 5600
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 2.7040764460709567e-05,
+      "loss": 0.2708,
+      "step": 5610
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 2.6989935956084173e-05,
+      "loss": 0.2596,
+      "step": 5620
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 2.693910745145878e-05,
+      "loss": 0.3078,
+      "step": 5630
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.6888278946833382e-05,
+      "loss": 0.2736,
+      "step": 5640
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.683745044220799e-05,
+      "loss": 0.2912,
+      "step": 5650
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.6786621937582595e-05,
+      "loss": 0.2675,
+      "step": 5660
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 2.67357934329572e-05,
+      "loss": 0.224,
+      "step": 5670
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 2.6684964928331808e-05,
+      "loss": 0.2435,
+      "step": 5680
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.6634136423706417e-05,
+      "loss": 0.2841,
+      "step": 5690
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.6583307919081023e-05,
+      "loss": 0.2968,
+      "step": 5700
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 2.653247941445563e-05,
+      "loss": 0.268,
+      "step": 5710
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.6481650909830236e-05,
+      "loss": 0.305,
+      "step": 5720
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 2.6430822405204842e-05,
+      "loss": 0.3122,
+      "step": 5730
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2.637999390057945e-05,
+      "loss": 0.2743,
+      "step": 5740
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 2.632916539595405e-05,
+      "loss": 0.2445,
+      "step": 5750
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.6278336891328658e-05,
+      "loss": 0.2433,
+      "step": 5760
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 2.6227508386703264e-05,
+      "loss": 0.292,
+      "step": 5770
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 2.617667988207787e-05,
+      "loss": 0.275,
+      "step": 5780
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 2.6125851377452477e-05,
+      "loss": 0.291,
+      "step": 5790
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 2.6075022872827083e-05,
+      "loss": 0.2679,
+      "step": 5800
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 2.6024194368201686e-05,
+      "loss": 0.2999,
+      "step": 5810
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 2.5973365863576292e-05,
+      "loss": 0.2667,
+      "step": 5820
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.59225373589509e-05,
+      "loss": 0.2802,
+      "step": 5830
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 2.5871708854325505e-05,
+      "loss": 0.2553,
+      "step": 5840
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 2.582088034970011e-05,
+      "loss": 0.2629,
+      "step": 5850
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 2.5770051845074717e-05,
+      "loss": 0.2526,
+      "step": 5860
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 2.5719223340449327e-05,
+      "loss": 0.2936,
+      "step": 5870
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 2.5668394835823933e-05,
+      "loss": 0.2666,
+      "step": 5880
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.561756633119854e-05,
+      "loss": 0.2775,
+      "step": 5890
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.5566737826573146e-05,
+      "loss": 0.3416,
+      "step": 5900
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.5515909321947752e-05,
+      "loss": 0.3037,
+      "step": 5910
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 2.5465080817322355e-05,
+      "loss": 0.2469,
+      "step": 5920
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 2.541425231269696e-05,
+      "loss": 0.2681,
+      "step": 5930
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 2.5363423808071568e-05,
+      "loss": 0.2763,
+      "step": 5940
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.5312595303446174e-05,
+      "loss": 0.2791,
+      "step": 5950
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 2.526176679882078e-05,
+      "loss": 0.2552,
+      "step": 5960
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 2.5210938294195386e-05,
+      "loss": 0.2521,
+      "step": 5970
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 2.5160109789569993e-05,
+      "loss": 0.2504,
+      "step": 5980
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 2.5109281284944596e-05,
+      "loss": 0.2719,
+      "step": 5990
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 2.5058452780319202e-05,
+      "loss": 0.2659,
+      "step": 6000
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 2.5007624275693808e-05,
+      "loss": 0.228,
+      "step": 6010
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.4956795771068418e-05,
+      "loss": 0.265,
+      "step": 6020
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 2.4905967266443024e-05,
+      "loss": 0.2356,
+      "step": 6030
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 2.4855138761817627e-05,
+      "loss": 0.2407,
+      "step": 6040
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.4804310257192233e-05,
+      "loss": 0.2858,
+      "step": 6050
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 2.475348175256684e-05,
+      "loss": 0.2798,
+      "step": 6060
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 2.4702653247941446e-05,
+      "loss": 0.2784,
+      "step": 6070
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 2.4651824743316052e-05,
+      "loss": 0.2752,
+      "step": 6080
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.460099623869066e-05,
+      "loss": 0.2951,
+      "step": 6090
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 2.4550167734065265e-05,
+      "loss": 0.2302,
+      "step": 6100
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 2.449933922943987e-05,
+      "loss": 0.2461,
+      "step": 6110
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.4448510724814477e-05,
+      "loss": 0.3365,
+      "step": 6120
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.4397682220189084e-05,
+      "loss": 0.2875,
+      "step": 6130
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 2.434685371556369e-05,
+      "loss": 0.2875,
+      "step": 6140
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 2.4296025210938296e-05,
+      "loss": 0.2289,
+      "step": 6150
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.42451967063129e-05,
+      "loss": 0.2926,
+      "step": 6160
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 2.4194368201687505e-05,
+      "loss": 0.2581,
+      "step": 6170
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 2.4143539697062115e-05,
+      "loss": 0.2636,
+      "step": 6180
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 2.409271119243672e-05,
+      "loss": 0.2825,
+      "step": 6190
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.4041882687811328e-05,
+      "loss": 0.2544,
+      "step": 6200
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 2.399105418318593e-05,
+      "loss": 0.2817,
+      "step": 6210
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 2.3940225678560537e-05,
+      "loss": 0.2869,
+      "step": 6220
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.3889397173935143e-05,
+      "loss": 0.2997,
+      "step": 6230
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 2.383856866930975e-05,
+      "loss": 0.314,
+      "step": 6240
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 2.3787740164684356e-05,
+      "loss": 0.2673,
+      "step": 6250
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 2.3736911660058962e-05,
+      "loss": 0.2711,
+      "step": 6260
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.368608315543357e-05,
+      "loss": 0.2742,
+      "step": 6270
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.3635254650808175e-05,
+      "loss": 0.269,
+      "step": 6280
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 2.358442614618278e-05,
+      "loss": 0.2551,
+      "step": 6290
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 2.3533597641557387e-05,
+      "loss": 0.3102,
+      "step": 6300
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2.3482769136931993e-05,
+      "loss": 0.2677,
+      "step": 6310
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 2.34319406323066e-05,
+      "loss": 0.2311,
+      "step": 6320
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.3381112127681203e-05,
+      "loss": 0.3033,
+      "step": 6330
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 2.333028362305581e-05,
+      "loss": 0.2558,
+      "step": 6340
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.327945511843042e-05,
+      "loss": 0.2468,
+      "step": 6350
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.3228626613805025e-05,
+      "loss": 0.3158,
+      "step": 6360
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 2.317779810917963e-05,
+      "loss": 0.3092,
+      "step": 6370
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.3126969604554234e-05,
+      "loss": 0.2386,
+      "step": 6380
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 2.307614109992884e-05,
+      "loss": 0.2975,
+      "step": 6390
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.3025312595303447e-05,
+      "loss": 0.2799,
+      "step": 6400
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 2.2974484090678053e-05,
+      "loss": 0.2583,
+      "step": 6410
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.292365558605266e-05,
+      "loss": 0.243,
+      "step": 6420
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 2.2872827081427266e-05,
+      "loss": 0.263,
+      "step": 6430
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 2.2821998576801872e-05,
+      "loss": 0.2433,
+      "step": 6440
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.2771170072176478e-05,
+      "loss": 0.2578,
+      "step": 6450
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 2.2720341567551084e-05,
+      "loss": 0.2388,
+      "step": 6460
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.266951306292569e-05,
+      "loss": 0.287,
+      "step": 6470
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.2618684558300297e-05,
+      "loss": 0.2906,
+      "step": 6480
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.2567856053674903e-05,
+      "loss": 0.2817,
+      "step": 6490
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 2.2517027549049506e-05,
+      "loss": 0.2926,
+      "step": 6500
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 2.2466199044424112e-05,
+      "loss": 0.222,
+      "step": 6510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.241537053979872e-05,
+      "loss": 0.2619,
+      "step": 6520
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 2.236454203517333e-05,
+      "loss": 0.2835,
+      "step": 6530
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 2.2313713530547935e-05,
+      "loss": 0.295,
+      "step": 6540
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 2.2262885025922538e-05,
+      "loss": 0.2495,
+      "step": 6550
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.2212056521297144e-05,
+      "loss": 0.2623,
+      "step": 6560
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9424666666666667,
+      "eval_loss": 0.15051521360874176,
+      "eval_runtime": 74.469,
+      "eval_samples_per_second": 402.852,
+      "eval_steps_per_second": 25.178,
+      "step": 6562
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 2.216122801667175e-05,
+      "loss": 0.2739,
+      "step": 6570
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 2.2110399512046356e-05,
+      "loss": 0.2467,
+      "step": 6580
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 2.2059571007420963e-05,
+      "loss": 0.2705,
+      "step": 6590
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 2.200874250279557e-05,
+      "loss": 0.2781,
+      "step": 6600
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 2.1957913998170175e-05,
+      "loss": 0.2114,
+      "step": 6610
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 2.190708549354478e-05,
+      "loss": 0.2453,
+      "step": 6620
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 2.1856256988919388e-05,
+      "loss": 0.2829,
+      "step": 6630
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 2.1805428484293994e-05,
+      "loss": 0.2577,
+      "step": 6640
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 2.17545999796686e-05,
+      "loss": 0.2706,
+      "step": 6650
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 2.1703771475043207e-05,
+      "loss": 0.2639,
+      "step": 6660
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 2.165294297041781e-05,
+      "loss": 0.295,
+      "step": 6670
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 2.1602114465792416e-05,
+      "loss": 0.2389,
+      "step": 6680
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 2.1551285961167022e-05,
+      "loss": 0.268,
+      "step": 6690
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.150045745654163e-05,
+      "loss": 0.2517,
+      "step": 6700
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.1449628951916238e-05,
+      "loss": 0.2653,
+      "step": 6710
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 2.139880044729084e-05,
+      "loss": 0.2543,
+      "step": 6720
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 2.1347971942665447e-05,
+      "loss": 0.268,
+      "step": 6730
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 2.1297143438040054e-05,
+      "loss": 0.3018,
+      "step": 6740
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 2.124631493341466e-05,
+      "loss": 0.2564,
+      "step": 6750
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 2.1195486428789266e-05,
+      "loss": 0.2719,
+      "step": 6760
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 2.1144657924163873e-05,
+      "loss": 0.2458,
+      "step": 6770
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 2.109382941953848e-05,
+      "loss": 0.2508,
+      "step": 6780
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 2.1043000914913082e-05,
+      "loss": 0.2818,
+      "step": 6790
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 2.099217241028769e-05,
+      "loss": 0.2574,
+      "step": 6800
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 2.0941343905662298e-05,
+      "loss": 0.2729,
+      "step": 6810
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 2.0890515401036904e-05,
+      "loss": 0.2722,
+      "step": 6820
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 2.083968689641151e-05,
+      "loss": 0.2729,
+      "step": 6830
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 2.0788858391786113e-05,
+      "loss": 0.2792,
+      "step": 6840
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 2.073802988716072e-05,
+      "loss": 0.2684,
+      "step": 6850
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 2.0687201382535326e-05,
+      "loss": 0.2739,
+      "step": 6860
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 2.0636372877909932e-05,
+      "loss": 0.2344,
+      "step": 6870
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 2.0585544373284538e-05,
+      "loss": 0.2771,
+      "step": 6880
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 2.0534715868659145e-05,
+      "loss": 0.2585,
+      "step": 6890
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 2.048388736403375e-05,
+      "loss": 0.291,
+      "step": 6900
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 2.0433058859408357e-05,
+      "loss": 0.27,
+      "step": 6910
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 2.0382230354782963e-05,
+      "loss": 0.2856,
+      "step": 6920
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 2.033140185015757e-05,
+      "loss": 0.2781,
+      "step": 6930
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 2.0280573345532176e-05,
+      "loss": 0.2722,
+      "step": 6940
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 2.0229744840906782e-05,
+      "loss": 0.273,
+      "step": 6950
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 2.0178916336281385e-05,
+      "loss": 0.2337,
+      "step": 6960
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 2.012808783165599e-05,
+      "loss": 0.263,
+      "step": 6970
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 2.00772593270306e-05,
+      "loss": 0.2523,
+      "step": 6980
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 2.0026430822405207e-05,
+      "loss": 0.2619,
+      "step": 6990
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 1.9975602317779814e-05,
+      "loss": 0.2895,
+      "step": 7000
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 1.9924773813154417e-05,
+      "loss": 0.2663,
+      "step": 7010
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 1.9873945308529023e-05,
+      "loss": 0.2589,
+      "step": 7020
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 1.982311680390363e-05,
+      "loss": 0.2609,
+      "step": 7030
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 1.9772288299278235e-05,
+      "loss": 0.2758,
+      "step": 7040
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 1.9721459794652842e-05,
+      "loss": 0.2392,
+      "step": 7050
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 1.9670631290027448e-05,
+      "loss": 0.2647,
+      "step": 7060
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 1.9619802785402054e-05,
+      "loss": 0.2922,
+      "step": 7070
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 1.956897428077666e-05,
+      "loss": 0.2438,
+      "step": 7080
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 1.9518145776151267e-05,
+      "loss": 0.2767,
+      "step": 7090
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 1.9467317271525873e-05,
+      "loss": 0.3042,
+      "step": 7100
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 1.941648876690048e-05,
+      "loss": 0.2368,
+      "step": 7110
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 1.9365660262275086e-05,
+      "loss": 0.2631,
+      "step": 7120
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 1.931483175764969e-05,
+      "loss": 0.2508,
+      "step": 7130
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 1.9264003253024295e-05,
+      "loss": 0.2957,
+      "step": 7140
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 1.9213174748398905e-05,
+      "loss": 0.2844,
+      "step": 7150
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 1.916234624377351e-05,
+      "loss": 0.2879,
+      "step": 7160
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.9111517739148117e-05,
+      "loss": 0.2603,
+      "step": 7170
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.906068923452272e-05,
+      "loss": 0.2599,
+      "step": 7180
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 1.9009860729897326e-05,
+      "loss": 0.2482,
+      "step": 7190
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 1.8959032225271933e-05,
+      "loss": 0.2922,
+      "step": 7200
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 1.890820372064654e-05,
+      "loss": 0.2772,
+      "step": 7210
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 1.8857375216021145e-05,
+      "loss": 0.2745,
+      "step": 7220
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 1.880654671139575e-05,
+      "loss": 0.2285,
+      "step": 7230
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 1.8755718206770358e-05,
+      "loss": 0.2539,
+      "step": 7240
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 1.8704889702144964e-05,
+      "loss": 0.2753,
+      "step": 7250
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 1.865406119751957e-05,
+      "loss": 0.3329,
+      "step": 7260
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.8603232692894177e-05,
+      "loss": 0.2594,
+      "step": 7270
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 1.8552404188268783e-05,
+      "loss": 0.2728,
+      "step": 7280
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 1.850157568364339e-05,
+      "loss": 0.2622,
+      "step": 7290
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 1.8450747179017992e-05,
+      "loss": 0.2486,
+      "step": 7300
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 1.83999186743926e-05,
+      "loss": 0.3036,
+      "step": 7310
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.8349090169767205e-05,
+      "loss": 0.2673,
+      "step": 7320
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 1.8298261665141814e-05,
+      "loss": 0.2592,
+      "step": 7330
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 1.824743316051642e-05,
+      "loss": 0.2545,
+      "step": 7340
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 1.8196604655891024e-05,
+      "loss": 0.2641,
+      "step": 7350
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 1.814577615126563e-05,
+      "loss": 0.2792,
+      "step": 7360
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 1.8094947646640236e-05,
+      "loss": 0.2633,
+      "step": 7370
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 1.8044119142014842e-05,
+      "loss": 0.2246,
+      "step": 7380
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 1.799329063738945e-05,
+      "loss": 0.2318,
+      "step": 7390
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 1.7942462132764055e-05,
+      "loss": 0.2581,
+      "step": 7400
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 1.789163362813866e-05,
+      "loss": 0.2702,
+      "step": 7410
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 1.7840805123513268e-05,
+      "loss": 0.2433,
+      "step": 7420
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 1.7789976618887874e-05,
+      "loss": 0.2504,
+      "step": 7430
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 1.773914811426248e-05,
+      "loss": 0.2565,
+      "step": 7440
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 1.7688319609637087e-05,
+      "loss": 0.2782,
+      "step": 7450
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 1.7637491105011693e-05,
+      "loss": 0.2342,
+      "step": 7460
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 1.7586662600386296e-05,
+      "loss": 0.2448,
+      "step": 7470
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 1.7535834095760902e-05,
+      "loss": 0.2472,
+      "step": 7480
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 1.7485005591135508e-05,
+      "loss": 0.2698,
+      "step": 7490
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 1.7434177086510115e-05,
+      "loss": 0.2724,
+      "step": 7500
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 1.7383348581884724e-05,
+      "loss": 0.2529,
+      "step": 7510
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.7332520077259327e-05,
+      "loss": 0.2571,
+      "step": 7520
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.7281691572633933e-05,
+      "loss": 0.2514,
+      "step": 7530
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 1.723086306800854e-05,
+      "loss": 0.2749,
+      "step": 7540
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 1.7180034563383146e-05,
+      "loss": 0.2109,
+      "step": 7550
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 1.7129206058757752e-05,
+      "loss": 0.245,
+      "step": 7560
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 1.707837755413236e-05,
+      "loss": 0.2869,
+      "step": 7570
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.7027549049506965e-05,
+      "loss": 0.2394,
+      "step": 7580
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 1.6976720544881568e-05,
+      "loss": 0.2881,
+      "step": 7590
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 1.6925892040256177e-05,
+      "loss": 0.2736,
+      "step": 7600
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 1.6875063535630784e-05,
+      "loss": 0.2115,
+      "step": 7610
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 1.682423503100539e-05,
+      "loss": 0.3126,
+      "step": 7620
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 1.6773406526379996e-05,
+      "loss": 0.297,
+      "step": 7630
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.67225780217546e-05,
+      "loss": 0.2825,
+      "step": 7640
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.6671749517129205e-05,
+      "loss": 0.2791,
+      "step": 7650
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9460333333333333,
+      "eval_loss": 0.14077039062976837,
+      "eval_runtime": 74.3692,
+      "eval_samples_per_second": 403.393,
+      "eval_steps_per_second": 25.212,
+      "step": 7656
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 1.6620921012503812e-05,
+      "loss": 0.2775,
+      "step": 7660
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 1.6570092507878418e-05,
+      "loss": 0.2384,
+      "step": 7670
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 1.6519264003253024e-05,
+      "loss": 0.2364,
+      "step": 7680
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 1.6468435498627634e-05,
+      "loss": 0.3096,
+      "step": 7690
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 1.6417606994002237e-05,
+      "loss": 0.2338,
+      "step": 7700
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.6366778489376843e-05,
+      "loss": 0.2629,
+      "step": 7710
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 1.631594998475145e-05,
+      "loss": 0.2577,
+      "step": 7720
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 1.6265121480126056e-05,
+      "loss": 0.3052,
+      "step": 7730
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 1.6214292975500662e-05,
+      "loss": 0.2709,
+      "step": 7740
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 1.616346447087527e-05,
+      "loss": 0.2972,
+      "step": 7750
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 1.611263596624987e-05,
+      "loss": 0.271,
+      "step": 7760
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 1.6061807461624478e-05,
+      "loss": 0.2501,
+      "step": 7770
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 1.6010978956999087e-05,
+      "loss": 0.2376,
+      "step": 7780
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 1.5960150452373694e-05,
+      "loss": 0.2411,
+      "step": 7790
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 1.59093219477483e-05,
+      "loss": 0.2698,
+      "step": 7800
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 1.5858493443122903e-05,
+      "loss": 0.2305,
+      "step": 7810
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 1.580766493849751e-05,
+      "loss": 0.2443,
+      "step": 7820
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 1.5756836433872115e-05,
+      "loss": 0.2704,
+      "step": 7830
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 1.570600792924672e-05,
+      "loss": 0.2962,
+      "step": 7840
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 1.5655179424621328e-05,
+      "loss": 0.2886,
+      "step": 7850
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 1.5604350919995934e-05,
+      "loss": 0.2766,
+      "step": 7860
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.555352241537054e-05,
+      "loss": 0.2416,
+      "step": 7870
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.5502693910745147e-05,
+      "loss": 0.2797,
+      "step": 7880
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.5451865406119753e-05,
+      "loss": 0.2345,
+      "step": 7890
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.540103690149436e-05,
+      "loss": 0.2667,
+      "step": 7900
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.5350208396868966e-05,
+      "loss": 0.2788,
+      "step": 7910
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.5299379892243572e-05,
+      "loss": 0.2653,
+      "step": 7920
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.5248551387618176e-05,
+      "loss": 0.2869,
+      "step": 7930
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 1.5197722882992783e-05,
+      "loss": 0.2389,
+      "step": 7940
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.5146894378367387e-05,
+      "loss": 0.2364,
+      "step": 7950
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.5096065873741997e-05,
+      "loss": 0.2789,
+      "step": 7960
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.5045237369116602e-05,
+      "loss": 0.2823,
+      "step": 7970
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.4994408864491208e-05,
+      "loss": 0.238,
+      "step": 7980
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.4943580359865814e-05,
+      "loss": 0.2849,
+      "step": 7990
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.4892751855240419e-05,
+      "loss": 0.2684,
+      "step": 8000
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 1.4841923350615025e-05,
+      "loss": 0.2838,
+      "step": 8010
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.4791094845989631e-05,
+      "loss": 0.236,
+      "step": 8020
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.4740266341364236e-05,
+      "loss": 0.255,
+      "step": 8030
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.4689437836738846e-05,
+      "loss": 0.2676,
+      "step": 8040
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.463860933211345e-05,
+      "loss": 0.2618,
+      "step": 8050
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.4587780827488056e-05,
+      "loss": 0.2351,
+      "step": 8060
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.4536952322862663e-05,
+      "loss": 0.2786,
+      "step": 8070
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.4486123818237269e-05,
+      "loss": 0.2837,
+      "step": 8080
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.4435295313611874e-05,
+      "loss": 0.2869,
+      "step": 8090
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.438446680898648e-05,
+      "loss": 0.252,
+      "step": 8100
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.4333638304361086e-05,
+      "loss": 0.2536,
+      "step": 8110
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.428280979973569e-05,
+      "loss": 0.2345,
+      "step": 8120
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.42319812951103e-05,
+      "loss": 0.2853,
+      "step": 8130
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.4181152790484905e-05,
+      "loss": 0.2321,
+      "step": 8140
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.4130324285859511e-05,
+      "loss": 0.2593,
+      "step": 8150
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.4079495781234118e-05,
+      "loss": 0.2381,
+      "step": 8160
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.4028667276608722e-05,
+      "loss": 0.2652,
+      "step": 8170
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.3977838771983329e-05,
+      "loss": 0.2248,
+      "step": 8180
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 1.3927010267357935e-05,
+      "loss": 0.2824,
+      "step": 8190
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.387618176273254e-05,
+      "loss": 0.2407,
+      "step": 8200
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.3825353258107146e-05,
+      "loss": 0.2495,
+      "step": 8210
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.3774524753481754e-05,
+      "loss": 0.2906,
+      "step": 8220
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.372369624885636e-05,
+      "loss": 0.2712,
+      "step": 8230
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.3672867744230966e-05,
+      "loss": 0.2459,
+      "step": 8240
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.3622039239605573e-05,
+      "loss": 0.2699,
+      "step": 8250
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.3571210734980177e-05,
+      "loss": 0.2773,
+      "step": 8260
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.3520382230354783e-05,
+      "loss": 0.2597,
+      "step": 8270
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.346955372572939e-05,
+      "loss": 0.2767,
+      "step": 8280
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 1.3418725221103994e-05,
+      "loss": 0.2555,
+      "step": 8290
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.33678967164786e-05,
+      "loss": 0.2549,
+      "step": 8300
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.3317068211853209e-05,
+      "loss": 0.2828,
+      "step": 8310
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.3266239707227815e-05,
+      "loss": 0.2558,
+      "step": 8320
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.3215411202602421e-05,
+      "loss": 0.3045,
+      "step": 8330
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.3164582697977026e-05,
+      "loss": 0.2953,
+      "step": 8340
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.3113754193351632e-05,
+      "loss": 0.2865,
+      "step": 8350
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.3062925688726238e-05,
+      "loss": 0.2458,
+      "step": 8360
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 1.3012097184100843e-05,
+      "loss": 0.261,
+      "step": 8370
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.296126867947545e-05,
+      "loss": 0.2757,
+      "step": 8380
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.2910440174850056e-05,
+      "loss": 0.2582,
+      "step": 8390
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.2859611670224663e-05,
+      "loss": 0.2673,
+      "step": 8400
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 1.280878316559927e-05,
+      "loss": 0.273,
+      "step": 8410
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.2757954660973876e-05,
+      "loss": 0.2523,
+      "step": 8420
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 1.270712615634848e-05,
+      "loss": 0.2321,
+      "step": 8430
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.2656297651723087e-05,
+      "loss": 0.2257,
+      "step": 8440
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.2605469147097693e-05,
+      "loss": 0.2783,
+      "step": 8450
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.2554640642472298e-05,
+      "loss": 0.2617,
+      "step": 8460
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.2503812137846904e-05,
+      "loss": 0.2364,
+      "step": 8470
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 1.2452983633221512e-05,
+      "loss": 0.2419,
+      "step": 8480
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.2402155128596117e-05,
+      "loss": 0.2022,
+      "step": 8490
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.2351326623970723e-05,
+      "loss": 0.2307,
+      "step": 8500
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.230049811934533e-05,
+      "loss": 0.3113,
+      "step": 8510
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.2249669614719936e-05,
+      "loss": 0.2401,
+      "step": 8520
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.2198841110094542e-05,
+      "loss": 0.2508,
+      "step": 8530
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.2148012605469148e-05,
+      "loss": 0.25,
+      "step": 8540
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.2097184100843753e-05,
+      "loss": 0.2408,
+      "step": 8550
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.204635559621836e-05,
+      "loss": 0.2401,
+      "step": 8560
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.1995527091592965e-05,
+      "loss": 0.2487,
+      "step": 8570
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.1944698586967572e-05,
+      "loss": 0.269,
+      "step": 8580
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.1893870082342178e-05,
+      "loss": 0.2503,
+      "step": 8590
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.1843041577716784e-05,
+      "loss": 0.2458,
+      "step": 8600
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.179221307309139e-05,
+      "loss": 0.2856,
+      "step": 8610
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.1741384568465997e-05,
+      "loss": 0.2054,
+      "step": 8620
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.1690556063840601e-05,
+      "loss": 0.2708,
+      "step": 8630
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.163972755921521e-05,
+      "loss": 0.2451,
+      "step": 8640
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.1588899054589816e-05,
+      "loss": 0.2317,
+      "step": 8650
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 1.153807054996442e-05,
+      "loss": 0.2589,
+      "step": 8660
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.1487242045339026e-05,
+      "loss": 0.271,
+      "step": 8670
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.1436413540713633e-05,
+      "loss": 0.271,
+      "step": 8680
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.1385585036088239e-05,
+      "loss": 0.2849,
+      "step": 8690
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.1334756531462845e-05,
+      "loss": 0.2618,
+      "step": 8700
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.1283928026837452e-05,
+      "loss": 0.2549,
+      "step": 8710
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.1233099522212056e-05,
+      "loss": 0.2741,
+      "step": 8720
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.1182271017586664e-05,
+      "loss": 0.2706,
+      "step": 8730
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 1.1131442512961269e-05,
+      "loss": 0.25,
+      "step": 8740
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.1080614008335875e-05,
+      "loss": 0.2686,
+      "step": 8750
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9433333333333334,
+      "eval_loss": 0.14896263182163239,
+      "eval_runtime": 73.0613,
+      "eval_samples_per_second": 410.614,
+      "eval_steps_per_second": 25.663,
+      "step": 8750
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 1.1029785503710481e-05,
+      "loss": 0.2602,
+      "step": 8760
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.0978956999085088e-05,
+      "loss": 0.2701,
+      "step": 8770
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.0928128494459694e-05,
+      "loss": 0.2109,
+      "step": 8780
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.08772999898343e-05,
+      "loss": 0.2733,
+      "step": 8790
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.0826471485208905e-05,
+      "loss": 0.2805,
+      "step": 8800
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.0775642980583511e-05,
+      "loss": 0.2777,
+      "step": 8810
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.0724814475958119e-05,
+      "loss": 0.2652,
+      "step": 8820
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.0673985971332724e-05,
+      "loss": 0.245,
+      "step": 8830
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.062315746670733e-05,
+      "loss": 0.2591,
+      "step": 8840
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.0572328962081936e-05,
+      "loss": 0.2194,
+      "step": 8850
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.0521500457456541e-05,
+      "loss": 0.2654,
+      "step": 8860
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.0470671952831149e-05,
+      "loss": 0.2256,
+      "step": 8870
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.0419843448205755e-05,
+      "loss": 0.224,
+      "step": 8880
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.036901494358036e-05,
+      "loss": 0.2523,
+      "step": 8890
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.0318186438954966e-05,
+      "loss": 0.2654,
+      "step": 8900
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.0267357934329572e-05,
+      "loss": 0.2086,
+      "step": 8910
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.0216529429704179e-05,
+      "loss": 0.255,
+      "step": 8920
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.0165700925078785e-05,
+      "loss": 0.2498,
+      "step": 8930
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.0114872420453391e-05,
+      "loss": 0.2631,
+      "step": 8940
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.0064043915827996e-05,
+      "loss": 0.2311,
+      "step": 8950
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.0013215411202604e-05,
+      "loss": 0.2775,
+      "step": 8960
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 9.962386906577208e-06,
+      "loss": 0.2433,
+      "step": 8970
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 9.911558401951815e-06,
+      "loss": 0.2504,
+      "step": 8980
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 9.860729897326421e-06,
+      "loss": 0.2399,
+      "step": 8990
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 9.809901392701027e-06,
+      "loss": 0.269,
+      "step": 9000
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 9.759072888075633e-06,
+      "loss": 0.2414,
+      "step": 9010
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 9.70824438345024e-06,
+      "loss": 0.251,
+      "step": 9020
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 9.657415878824844e-06,
+      "loss": 0.2814,
+      "step": 9030
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 9.606587374199452e-06,
+      "loss": 0.2526,
+      "step": 9040
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 9.555758869574059e-06,
+      "loss": 0.2531,
+      "step": 9050
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 9.504930364948663e-06,
+      "loss": 0.2857,
+      "step": 9060
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 9.45410186032327e-06,
+      "loss": 0.2538,
+      "step": 9070
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 9.403273355697876e-06,
+      "loss": 0.2169,
+      "step": 9080
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 9.352444851072482e-06,
+      "loss": 0.2504,
+      "step": 9090
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 9.301616346447088e-06,
+      "loss": 0.2592,
+      "step": 9100
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 9.250787841821695e-06,
+      "loss": 0.2198,
+      "step": 9110
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 9.1999593371963e-06,
+      "loss": 0.218,
+      "step": 9120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 9.149130832570907e-06,
+      "loss": 0.2448,
+      "step": 9130
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 9.098302327945512e-06,
+      "loss": 0.262,
+      "step": 9140
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 9.047473823320118e-06,
+      "loss": 0.2334,
+      "step": 9150
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 8.996645318694724e-06,
+      "loss": 0.2356,
+      "step": 9160
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 8.94581681406933e-06,
+      "loss": 0.2476,
+      "step": 9170
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 8.894988309443937e-06,
+      "loss": 0.272,
+      "step": 9180
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 8.844159804818543e-06,
+      "loss": 0.3092,
+      "step": 9190
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 8.793331300193148e-06,
+      "loss": 0.2359,
+      "step": 9200
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 8.742502795567754e-06,
+      "loss": 0.2856,
+      "step": 9210
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 8.691674290942362e-06,
+      "loss": 0.2437,
+      "step": 9220
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 8.640845786316967e-06,
+      "loss": 0.2398,
+      "step": 9230
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 8.590017281691573e-06,
+      "loss": 0.2608,
+      "step": 9240
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 8.53918877706618e-06,
+      "loss": 0.2407,
+      "step": 9250
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 8.488360272440784e-06,
+      "loss": 0.2508,
+      "step": 9260
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 8.437531767815392e-06,
+      "loss": 0.2324,
+      "step": 9270
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 8.386703263189998e-06,
+      "loss": 0.2418,
+      "step": 9280
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 8.335874758564603e-06,
+      "loss": 0.2513,
+      "step": 9290
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 8.285046253939209e-06,
+      "loss": 0.3118,
+      "step": 9300
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 8.234217749313817e-06,
+      "loss": 0.2794,
+      "step": 9310
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 8.183389244688422e-06,
+      "loss": 0.2507,
+      "step": 9320
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 8.132560740063028e-06,
+      "loss": 0.2615,
+      "step": 9330
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 8.081732235437634e-06,
+      "loss": 0.2758,
+      "step": 9340
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 8.030903730812239e-06,
+      "loss": 0.2554,
+      "step": 9350
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 7.980075226186847e-06,
+      "loss": 0.2548,
+      "step": 9360
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 7.929246721561451e-06,
+      "loss": 0.2935,
+      "step": 9370
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 7.878418216936058e-06,
+      "loss": 0.2809,
+      "step": 9380
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 7.827589712310664e-06,
+      "loss": 0.2518,
+      "step": 9390
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 7.77676120768527e-06,
+      "loss": 0.2797,
+      "step": 9400
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 7.725932703059876e-06,
+      "loss": 0.2917,
+      "step": 9410
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 7.675104198434483e-06,
+      "loss": 0.2476,
+      "step": 9420
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 7.624275693809088e-06,
+      "loss": 0.2468,
+      "step": 9430
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 7.573447189183694e-06,
+      "loss": 0.2815,
+      "step": 9440
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 7.522618684558301e-06,
+      "loss": 0.2351,
+      "step": 9450
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 7.471790179932907e-06,
+      "loss": 0.2137,
+      "step": 9460
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 7.4209616753075125e-06,
+      "loss": 0.2452,
+      "step": 9470
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 7.370133170682118e-06,
+      "loss": 0.2384,
+      "step": 9480
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 7.319304666056725e-06,
+      "loss": 0.2246,
+      "step": 9490
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 7.268476161431331e-06,
+      "loss": 0.2446,
+      "step": 9500
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 7.217647656805937e-06,
+      "loss": 0.254,
+      "step": 9510
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 7.166819152180543e-06,
+      "loss": 0.2559,
+      "step": 9520
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 7.11599064755515e-06,
+      "loss": 0.2705,
+      "step": 9530
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 7.065162142929756e-06,
+      "loss": 0.2377,
+      "step": 9540
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 7.014333638304361e-06,
+      "loss": 0.2542,
+      "step": 9550
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 6.963505133678967e-06,
+      "loss": 0.2811,
+      "step": 9560
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 6.912676629053573e-06,
+      "loss": 0.2271,
+      "step": 9570
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 6.86184812442818e-06,
+      "loss": 0.2641,
+      "step": 9580
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 6.811019619802786e-06,
+      "loss": 0.2536,
+      "step": 9590
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 6.760191115177392e-06,
+      "loss": 0.2338,
+      "step": 9600
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 6.709362610551997e-06,
+      "loss": 0.273,
+      "step": 9610
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 6.658534105926604e-06,
+      "loss": 0.2285,
+      "step": 9620
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 6.607705601301211e-06,
+      "loss": 0.226,
+      "step": 9630
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 6.556877096675816e-06,
+      "loss": 0.2707,
+      "step": 9640
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 6.5060485920504215e-06,
+      "loss": 0.2216,
+      "step": 9650
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 6.455220087425028e-06,
+      "loss": 0.2863,
+      "step": 9660
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.404391582799635e-06,
+      "loss": 0.2327,
+      "step": 9670
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.35356307817424e-06,
+      "loss": 0.2231,
+      "step": 9680
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.302734573548847e-06,
+      "loss": 0.2226,
+      "step": 9690
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.251906068923452e-06,
+      "loss": 0.2448,
+      "step": 9700
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.201077564298058e-06,
+      "loss": 0.2476,
+      "step": 9710
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.150249059672665e-06,
+      "loss": 0.285,
+      "step": 9720
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.099420555047271e-06,
+      "loss": 0.2364,
+      "step": 9730
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.048592050421876e-06,
+      "loss": 0.2201,
+      "step": 9740
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 5.997763545796483e-06,
+      "loss": 0.2542,
+      "step": 9750
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 5.946935041171089e-06,
+      "loss": 0.2831,
+      "step": 9760
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.896106536545695e-06,
+      "loss": 0.2245,
+      "step": 9770
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.845278031920301e-06,
+      "loss": 0.221,
+      "step": 9780
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.794449527294908e-06,
+      "loss": 0.2711,
+      "step": 9790
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.743621022669513e-06,
+      "loss": 0.2221,
+      "step": 9800
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.6927925180441195e-06,
+      "loss": 0.2714,
+      "step": 9810
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.641964013418726e-06,
+      "loss": 0.2554,
+      "step": 9820
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.591135508793332e-06,
+      "loss": 0.2327,
+      "step": 9830
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.5403070041679375e-06,
+      "loss": 0.2219,
+      "step": 9840
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9445333333333333,
+      "eval_loss": 0.1479150652885437,
+      "eval_runtime": 72.9207,
+      "eval_samples_per_second": 411.406,
+      "eval_steps_per_second": 25.713,
+      "step": 9843
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.489478499542544e-06,
+      "loss": 0.2616,
+      "step": 9850
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.43864999491715e-06,
+      "loss": 0.2332,
+      "step": 9860
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 5.3878214902917556e-06,
+      "loss": 0.2644,
+      "step": 9870
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 5.336992985666362e-06,
+      "loss": 0.2248,
+      "step": 9880
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 5.286164481040968e-06,
+      "loss": 0.2342,
+      "step": 9890
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 5.235335976415574e-06,
+      "loss": 0.2636,
+      "step": 9900
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 5.18450747179018e-06,
+      "loss": 0.2274,
+      "step": 9910
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 5.133678967164786e-06,
+      "loss": 0.2569,
+      "step": 9920
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 5.0828504625393924e-06,
+      "loss": 0.257,
+      "step": 9930
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 5.032021957913998e-06,
+      "loss": 0.2574,
+      "step": 9940
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.981193453288604e-06,
+      "loss": 0.2497,
+      "step": 9950
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.9303649486632104e-06,
+      "loss": 0.2843,
+      "step": 9960
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.879536444037817e-06,
+      "loss": 0.2339,
+      "step": 9970
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.828707939412422e-06,
+      "loss": 0.2714,
+      "step": 9980
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.777879434787029e-06,
+      "loss": 0.2421,
+      "step": 9990
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.727050930161635e-06,
+      "loss": 0.226,
+      "step": 10000
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.676222425536241e-06,
+      "loss": 0.2154,
+      "step": 10010
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.625393920910847e-06,
+      "loss": 0.271,
+      "step": 10020
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.574565416285454e-06,
+      "loss": 0.2686,
+      "step": 10030
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.523736911660059e-06,
+      "loss": 0.241,
+      "step": 10040
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.472908407034665e-06,
+      "loss": 0.2343,
+      "step": 10050
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.422079902409272e-06,
+      "loss": 0.2375,
+      "step": 10060
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.371251397783877e-06,
+      "loss": 0.2483,
+      "step": 10070
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 4.320422893158483e-06,
+      "loss": 0.255,
+      "step": 10080
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.26959438853309e-06,
+      "loss": 0.2466,
+      "step": 10090
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.218765883907696e-06,
+      "loss": 0.2125,
+      "step": 10100
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 4.167937379282301e-06,
+      "loss": 0.2642,
+      "step": 10110
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.1171088746569085e-06,
+      "loss": 0.2515,
+      "step": 10120
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 4.066280370031514e-06,
+      "loss": 0.2595,
+      "step": 10130
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.015451865406119e-06,
+      "loss": 0.2453,
+      "step": 10140
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 3.964623360780726e-06,
+      "loss": 0.2294,
+      "step": 10150
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.913794856155332e-06,
+      "loss": 0.2307,
+      "step": 10160
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 3.862966351529938e-06,
+      "loss": 0.2736,
+      "step": 10170
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 3.812137846904544e-06,
+      "loss": 0.2634,
+      "step": 10180
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 3.7613093422791504e-06,
+      "loss": 0.2194,
+      "step": 10190
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.7104808376537563e-06,
+      "loss": 0.2694,
+      "step": 10200
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.6596523330283626e-06,
+      "loss": 0.2352,
+      "step": 10210
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 3.6088238284029684e-06,
+      "loss": 0.2582,
+      "step": 10220
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 3.557995323777575e-06,
+      "loss": 0.2618,
+      "step": 10230
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 3.5071668191521806e-06,
+      "loss": 0.2831,
+      "step": 10240
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.4563383145267864e-06,
+      "loss": 0.2546,
+      "step": 10250
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 3.405509809901393e-06,
+      "loss": 0.2658,
+      "step": 10260
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 3.3546813052759986e-06,
+      "loss": 0.2422,
+      "step": 10270
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 3.3038528006506053e-06,
+      "loss": 0.2447,
+      "step": 10280
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.2530242960252107e-06,
+      "loss": 0.2488,
+      "step": 10290
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 3.2021957913998174e-06,
+      "loss": 0.239,
+      "step": 10300
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 3.1513672867744233e-06,
+      "loss": 0.2494,
+      "step": 10310
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 3.100538782149029e-06,
+      "loss": 0.2396,
+      "step": 10320
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 3.0497102775236355e-06,
+      "loss": 0.2459,
+      "step": 10330
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 2.9988817728982413e-06,
+      "loss": 0.2494,
+      "step": 10340
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 2.9480532682728476e-06,
+      "loss": 0.2396,
+      "step": 10350
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 2.897224763647454e-06,
+      "loss": 0.2642,
+      "step": 10360
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 2.8463962590220598e-06,
+      "loss": 0.2628,
+      "step": 10370
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 2.795567754396666e-06,
+      "loss": 0.2478,
+      "step": 10380
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 2.744739249771272e-06,
+      "loss": 0.2648,
+      "step": 10390
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 2.6939107451458778e-06,
+      "loss": 0.2511,
+      "step": 10400
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 2.643082240520484e-06,
+      "loss": 0.2602,
+      "step": 10410
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 2.59225373589509e-06,
+      "loss": 0.2747,
+      "step": 10420
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 2.5414252312696962e-06,
+      "loss": 0.2353,
+      "step": 10430
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 2.490596726644302e-06,
+      "loss": 0.2455,
+      "step": 10440
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 2.4397682220189084e-06,
+      "loss": 0.2147,
+      "step": 10450
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 2.3889397173935147e-06,
+      "loss": 0.2492,
+      "step": 10460
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 2.3381112127681205e-06,
+      "loss": 0.223,
+      "step": 10470
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 2.287282708142727e-06,
+      "loss": 0.2466,
+      "step": 10480
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 2.2364542035173327e-06,
+      "loss": 0.2345,
+      "step": 10490
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 2.1856256988919385e-06,
+      "loss": 0.2591,
+      "step": 10500
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 2.134797194266545e-06,
+      "loss": 0.2531,
+      "step": 10510
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 2.0839686896411507e-06,
+      "loss": 0.273,
+      "step": 10520
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 2.033140185015757e-06,
+      "loss": 0.2249,
+      "step": 10530
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 1.982311680390363e-06,
+      "loss": 0.2338,
+      "step": 10540
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 1.931483175764969e-06,
+      "loss": 0.2881,
+      "step": 10550
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 1.8806546711395752e-06,
+      "loss": 0.212,
+      "step": 10560
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 1.8298261665141813e-06,
+      "loss": 0.2209,
+      "step": 10570
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 1.7789976618887876e-06,
+      "loss": 0.2659,
+      "step": 10580
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.7281691572633932e-06,
+      "loss": 0.2236,
+      "step": 10590
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 1.6773406526379993e-06,
+      "loss": 0.25,
+      "step": 10600
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 1.6265121480126054e-06,
+      "loss": 0.2704,
+      "step": 10610
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 1.5756836433872117e-06,
+      "loss": 0.2703,
+      "step": 10620
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 1.5248551387618177e-06,
+      "loss": 0.2542,
+      "step": 10630
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 1.4740266341364238e-06,
+      "loss": 0.2466,
+      "step": 10640
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 1.4231981295110299e-06,
+      "loss": 0.2301,
+      "step": 10650
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 1.372369624885636e-06,
+      "loss": 0.265,
+      "step": 10660
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 1.321541120260242e-06,
+      "loss": 0.2369,
+      "step": 10670
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 1.2707126156348481e-06,
+      "loss": 0.2566,
+      "step": 10680
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 1.2198841110094542e-06,
+      "loss": 0.2768,
+      "step": 10690
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 1.1690556063840603e-06,
+      "loss": 0.286,
+      "step": 10700
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 1.1182271017586663e-06,
+      "loss": 0.2239,
+      "step": 10710
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 1.0673985971332724e-06,
+      "loss": 0.2328,
+      "step": 10720
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.0165700925078785e-06,
+      "loss": 0.2134,
+      "step": 10730
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 9.657415878824846e-07,
+      "loss": 0.2364,
+      "step": 10740
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 9.149130832570906e-07,
+      "loss": 0.2947,
+      "step": 10750
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 8.640845786316966e-07,
+      "loss": 0.2441,
+      "step": 10760
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 8.132560740063027e-07,
+      "loss": 0.1992,
+      "step": 10770
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 7.624275693809089e-07,
+      "loss": 0.2691,
+      "step": 10780
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 7.115990647555149e-07,
+      "loss": 0.2536,
+      "step": 10790
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 6.60770560130121e-07,
+      "loss": 0.2289,
+      "step": 10800
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 6.099420555047271e-07,
+      "loss": 0.2571,
+      "step": 10810
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 5.591135508793332e-07,
+      "loss": 0.236,
+      "step": 10820
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.082850462539392e-07,
+      "loss": 0.2579,
+      "step": 10830
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.574565416285453e-07,
+      "loss": 0.2325,
+      "step": 10840
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.0662803700315134e-07,
+      "loss": 0.2756,
+      "step": 10850
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 3.5579953237775747e-07,
+      "loss": 0.272,
+      "step": 10860
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 3.0497102775236355e-07,
+      "loss": 0.2585,
+      "step": 10870
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 2.541425231269696e-07,
+      "loss": 0.2585,
+      "step": 10880
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 2.0331401850157567e-07,
+      "loss": 0.243,
+      "step": 10890
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 1.5248551387618177e-07,
+      "loss": 0.2529,
+      "step": 10900
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 1.0165700925078784e-07,
+      "loss": 0.2215,
+      "step": 10910
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 5.082850462539392e-08,
+      "loss": 0.2311,
+      "step": 10920
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 0.0,
+      "loss": 0.2552,
+      "step": 10930
+    },
+    {
+      "epoch": 9.99,
+      "eval_accuracy": 0.9506666666666667,
+      "eval_loss": 0.1330413967370987,
+      "eval_runtime": 72.7989,
+      "eval_samples_per_second": 412.094,
+      "eval_steps_per_second": 25.756,
+      "step": 10930
+    },
+    {
+      "epoch": 9.99,
+      "step": 10930,
+      "total_flos": 1.4854589405842637e+19,
+      "train_loss": 0.31431262887952965,
+      "train_runtime": 4384.3481,
+      "train_samples_per_second": 159.659,
+      "train_steps_per_second": 2.493
+    }
+  ],
+  "max_steps": 10930,
+  "num_train_epochs": 10,
+  "total_flos": 1.4854589405842637e+19,
+  "trial_name": null,
+  "trial_params": null
+}