diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4876 @@
+{
+  "best_metric": 0.8974400257607471,
+  "best_model_checkpoint": "/home/cloudwalker/ASVmodel/wavlm-base/checkpoint-793",
+  "epoch": 9.996848408446265,
+  "eval_steps": 500,
+  "global_step": 7930,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.783102143757881e-06,
+      "loss": 0.707,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 7.566204287515762e-06,
+      "loss": 0.6657,
+      "step": 20
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.1349306431273643e-05,
+      "loss": 0.5764,
+      "step": 30
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.5132408575031524e-05,
+      "loss": 0.5123,
+      "step": 40
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.8915510718789407e-05,
+      "loss": 0.3307,
+      "step": 50
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.2698612862547287e-05,
+      "loss": 0.3335,
+      "step": 60
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.648171500630517e-05,
+      "loss": 0.2621,
+      "step": 70
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 3.026481715006305e-05,
+      "loss": 0.2639,
+      "step": 80
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 3.404791929382093e-05,
+      "loss": 0.1646,
+      "step": 90
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 3.7831021437578815e-05,
+      "loss": 0.1115,
+      "step": 100
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.161412358133669e-05,
+      "loss": 0.1056,
+      "step": 110
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.5397225725094573e-05,
+      "loss": 0.151,
+      "step": 120
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.918032786885245e-05,
+      "loss": 0.0505,
+      "step": 130
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.296343001261034e-05,
+      "loss": 0.1391,
+      "step": 140
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.6746532156368215e-05,
+      "loss": 0.0993,
+      "step": 150
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 6.05296343001261e-05,
+      "loss": 0.1814,
+      "step": 160
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 6.431273644388397e-05,
+      "loss": 0.092,
+      "step": 170
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 6.809583858764186e-05,
+      "loss": 0.0772,
+      "step": 180
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 7.187894073139974e-05,
+      "loss": 0.0281,
+      "step": 190
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 7.566204287515763e-05,
+      "loss": 0.0335,
+      "step": 200
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 7.944514501891549e-05,
+      "loss": 0.0062,
+      "step": 210
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 8.322824716267338e-05,
+      "loss": 0.0222,
+      "step": 220
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 8.701134930643127e-05,
+      "loss": 0.0996,
+      "step": 230
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.079445145018915e-05,
+      "loss": 0.0351,
+      "step": 240
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.457755359394704e-05,
+      "loss": 0.2802,
+      "step": 250
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.83606557377049e-05,
+      "loss": 0.1714,
+      "step": 260
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00010214375788146279,
+      "loss": 0.0,
+      "step": 270
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00010592686002522068,
+      "loss": 0.0,
+      "step": 280
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00010970996216897855,
+      "loss": 0.4595,
+      "step": 290
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00011349306431273643,
+      "loss": 0.2479,
+      "step": 300
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001172761664564943,
+      "loss": 0.332,
+      "step": 310
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001210592686002522,
+      "loss": 0.0512,
+      "step": 320
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00012484237074401007,
+      "loss": 0.0,
+      "step": 330
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00012862547288776795,
+      "loss": 0.9099,
+      "step": 340
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00013240857503152585,
+      "loss": 0.4185,
+      "step": 350
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00013619167717528373,
+      "loss": 0.407,
+      "step": 360
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001399747793190416,
+      "loss": 0.3369,
+      "step": 370
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00014375788146279948,
+      "loss": 0.35,
+      "step": 380
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00014754098360655736,
+      "loss": 0.3807,
+      "step": 390
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00015132408575031526,
+      "loss": 0.3404,
+      "step": 400
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.00015510718789407313,
+      "loss": 0.2999,
+      "step": 410
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00015889029003783098,
+      "loss": 0.3987,
+      "step": 420
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001626733921815889,
+      "loss": 0.3099,
+      "step": 430
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.00016645649432534676,
+      "loss": 0.343,
+      "step": 440
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.00017023959646910467,
+      "loss": 0.3723,
+      "step": 450
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.00017402269861286254,
+      "loss": 0.314,
+      "step": 460
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.0001778058007566204,
+      "loss": 0.3544,
+      "step": 470
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001815889029003783,
+      "loss": 0.3511,
+      "step": 480
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00018537200504413617,
+      "loss": 0.3602,
+      "step": 490
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.00018915510718789407,
+      "loss": 0.3795,
+      "step": 500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019293820933165195,
+      "loss": 0.4271,
+      "step": 510
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001967213114754098,
+      "loss": 0.3461,
+      "step": 520
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.0002005044136191677,
+      "loss": 0.3774,
+      "step": 530
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.00020428751576292558,
+      "loss": 0.3948,
+      "step": 540
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00020807061790668348,
+      "loss": 0.3301,
+      "step": 550
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00021185372005044136,
+      "loss": 0.3641,
+      "step": 560
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0002156368221941992,
+      "loss": 0.3732,
+      "step": 570
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0002194199243379571,
+      "loss": 0.2903,
+      "step": 580
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.00022320302648171498,
+      "loss": 0.3845,
+      "step": 590
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.00022698612862547286,
+      "loss": 0.298,
+      "step": 600
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.00023076923076923076,
+      "loss": 0.3758,
+      "step": 610
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0002345523329129886,
+      "loss": 0.2544,
+      "step": 620
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.00023833543505674652,
+      "loss": 0.2796,
+      "step": 630
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.0002421185372005044,
+      "loss": 0.3833,
+      "step": 640
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.0002459016393442623,
+      "loss": 0.3147,
+      "step": 650
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.00024968474148802014,
+      "loss": 0.3005,
+      "step": 660
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.00025346784363177805,
+      "loss": 0.2923,
+      "step": 670
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.0002572509457755359,
+      "loss": 0.3583,
+      "step": 680
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.0002610340479192938,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.0002648171500630517,
+      "loss": 0.3117,
+      "step": 700
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.00026860025220680955,
+      "loss": 0.3323,
+      "step": 710
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00027238335435056745,
+      "loss": 0.4065,
+      "step": 720
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0002761664564943253,
+      "loss": 0.3212,
+      "step": 730
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0002799495586380832,
+      "loss": 0.2834,
+      "step": 740
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.0002837326607818411,
+      "loss": 0.3154,
+      "step": 750
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00028751576292559896,
+      "loss": 0.3027,
+      "step": 760
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.00029129886506935686,
+      "loss": 0.3909,
+      "step": 770
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0002950819672131147,
+      "loss": 0.3255,
+      "step": 780
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0002988650693568726,
+      "loss": 0.3744,
+      "step": 790
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.3307398855686188,
+      "eval_runtime": 507.4905,
+      "eval_samples_per_second": 48.955,
+      "eval_steps_per_second": 24.477,
+      "step": 793
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.00029970575872215215,
+      "loss": 0.3253,
+      "step": 800
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00029928541403951236,
+      "loss": 0.301,
+      "step": 810
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.0002988650693568726,
+      "loss": 0.3422,
+      "step": 820
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.00029844472467423287,
+      "loss": 0.3214,
+      "step": 830
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0002980243799915931,
+      "loss": 0.3457,
+      "step": 840
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00029760403530895333,
+      "loss": 0.325,
+      "step": 850
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.00029718369062631353,
+      "loss": 0.378,
+      "step": 860
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0002967633459436738,
+      "loss": 0.3612,
+      "step": 870
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.000296343001261034,
+      "loss": 0.2886,
+      "step": 880
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.00029592265657839425,
+      "loss": 0.3075,
+      "step": 890
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0002955023118957545,
+      "loss": 0.2418,
+      "step": 900
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.0002950819672131147,
+      "loss": 0.4868,
+      "step": 910
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00029466162253047497,
+      "loss": 0.3235,
+      "step": 920
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.00029424127784783517,
+      "loss": 0.344,
+      "step": 930
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.00029382093316519543,
+      "loss": 0.4043,
+      "step": 940
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00029340058848255563,
+      "loss": 0.3372,
+      "step": 950
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.0002929802437999159,
+      "loss": 0.3483,
+      "step": 960
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.00029255989911727614,
+      "loss": 0.3533,
+      "step": 970
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.00029213955443463635,
+      "loss": 0.3612,
+      "step": 980
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.0002917192097519966,
+      "loss": 0.3848,
+      "step": 990
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.00029129886506935686,
+      "loss": 0.302,
+      "step": 1000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00029087852038671706,
+      "loss": 0.3491,
+      "step": 1010
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0002904581757040773,
+      "loss": 0.3638,
+      "step": 1020
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0002900378310214375,
+      "loss": 0.3399,
+      "step": 1030
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0002896174863387978,
+      "loss": 0.285,
+      "step": 1040
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.00028919714165615804,
+      "loss": 0.3085,
+      "step": 1050
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00028877679697351824,
+      "loss": 0.2929,
+      "step": 1060
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.0002883564522908785,
+      "loss": 0.277,
+      "step": 1070
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.00028793610760823876,
+      "loss": 0.3569,
+      "step": 1080
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.00028751576292559896,
+      "loss": 0.3388,
+      "step": 1090
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0002870954182429592,
+      "loss": 0.3657,
+      "step": 1100
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00028667507356031947,
+      "loss": 0.3547,
+      "step": 1110
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.0002862547288776797,
+      "loss": 0.3042,
+      "step": 1120
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00028583438419503993,
+      "loss": 0.2361,
+      "step": 1130
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.00028541403951240014,
+      "loss": 0.2841,
+      "step": 1140
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0002849936948297604,
+      "loss": 0.3706,
+      "step": 1150
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.00028457335014712065,
+      "loss": 0.325,
+      "step": 1160
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00028415300546448085,
+      "loss": 0.3592,
+      "step": 1170
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.0002837326607818411,
+      "loss": 0.2979,
+      "step": 1180
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0002833123160992013,
+      "loss": 0.4551,
+      "step": 1190
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00028289197141656157,
+      "loss": 0.2881,
+      "step": 1200
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.0002824716267339218,
+      "loss": 0.2841,
+      "step": 1210
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.00028205128205128203,
+      "loss": 0.2883,
+      "step": 1220
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0002816309373686423,
+      "loss": 0.3121,
+      "step": 1230
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0002812105926860025,
+      "loss": 0.3184,
+      "step": 1240
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.00028079024800336275,
+      "loss": 0.3308,
+      "step": 1250
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.00028036990332072295,
+      "loss": 0.3339,
+      "step": 1260
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.0002799495586380832,
+      "loss": 0.3412,
+      "step": 1270
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.00027952921395544346,
+      "loss": 0.3928,
+      "step": 1280
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.00027910886927280367,
+      "loss": 0.3866,
+      "step": 1290
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0002786885245901639,
+      "loss": 0.2891,
+      "step": 1300
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.0002782681799075241,
+      "loss": 0.3844,
+      "step": 1310
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.0002778478352248844,
+      "loss": 0.3529,
+      "step": 1320
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.0002774274905422446,
+      "loss": 0.3359,
+      "step": 1330
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.00027700714585960484,
+      "loss": 0.3122,
+      "step": 1340
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0002765868011769651,
+      "loss": 0.278,
+      "step": 1350
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0002761664564943253,
+      "loss": 0.3194,
+      "step": 1360
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00027574611181168556,
+      "loss": 0.361,
+      "step": 1370
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.00027532576712904576,
+      "loss": 0.3597,
+      "step": 1380
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.000274905422446406,
+      "loss": 0.3608,
+      "step": 1390
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.0002744850777637662,
+      "loss": 0.3089,
+      "step": 1400
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0002740647330811265,
+      "loss": 0.2781,
+      "step": 1410
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.00027364438839848674,
+      "loss": 0.3536,
+      "step": 1420
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.00027322404371584694,
+      "loss": 0.3071,
+      "step": 1430
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.0002728036990332072,
+      "loss": 0.358,
+      "step": 1440
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.00027238335435056745,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.00027196300966792766,
+      "loss": 0.406,
+      "step": 1460
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0002715426649852879,
+      "loss": 0.329,
+      "step": 1470
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.0002711223203026481,
+      "loss": 0.3082,
+      "step": 1480
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.0002707019756200084,
+      "loss": 0.3989,
+      "step": 1490
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00027028163093736863,
+      "loss": 0.2841,
+      "step": 1500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.00026986128625472883,
+      "loss": 0.282,
+      "step": 1510
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0002694409415720891,
+      "loss": 0.2896,
+      "step": 1520
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00026902059688944935,
+      "loss": 0.3382,
+      "step": 1530
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.00026860025220680955,
+      "loss": 0.2924,
+      "step": 1540
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 0.0002681799075241698,
+      "loss": 0.2875,
+      "step": 1550
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 0.00026775956284153007,
+      "loss": 0.3268,
+      "step": 1560
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 0.00026733921815889027,
+      "loss": 0.4224,
+      "step": 1570
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 0.0002669188734762505,
+      "loss": 0.3699,
+      "step": 1580
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.33422285318374634,
+      "eval_runtime": 509.6207,
+      "eval_samples_per_second": 48.75,
+      "eval_steps_per_second": 24.375,
+      "step": 1586
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.00026649852879361073,
+      "loss": 0.3175,
+      "step": 1590
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 0.000266078184110971,
+      "loss": 0.28,
+      "step": 1600
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 0.00026565783942833124,
+      "loss": 0.42,
+      "step": 1610
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 0.00026523749474569145,
+      "loss": 0.2945,
+      "step": 1620
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 0.0002648171500630517,
+      "loss": 0.3134,
+      "step": 1630
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 0.0002643968053804119,
+      "loss": 0.3812,
+      "step": 1640
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 0.00026397646069777216,
+      "loss": 0.3121,
+      "step": 1650
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 0.0002635561160151324,
+      "loss": 0.2884,
+      "step": 1660
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 0.0002631357713324926,
+      "loss": 0.3631,
+      "step": 1670
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 0.0002627154266498529,
+      "loss": 0.3598,
+      "step": 1680
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.0002622950819672131,
+      "loss": 0.3645,
+      "step": 1690
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 0.00026187473728457334,
+      "loss": 0.2908,
+      "step": 1700
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 0.00026145439260193354,
+      "loss": 0.3942,
+      "step": 1710
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 0.0002610340479192938,
+      "loss": 0.324,
+      "step": 1720
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 0.00026061370323665406,
+      "loss": 0.409,
+      "step": 1730
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 0.00026019335855401426,
+      "loss": 0.3598,
+      "step": 1740
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 0.0002597730138713745,
+      "loss": 0.3164,
+      "step": 1750
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 0.0002593526691887347,
+      "loss": 0.3452,
+      "step": 1760
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 0.000258932324506095,
+      "loss": 0.3503,
+      "step": 1770
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 0.0002585119798234552,
+      "loss": 0.3253,
+      "step": 1780
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 0.00025809163514081544,
+      "loss": 0.3734,
+      "step": 1790
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 0.0002576712904581757,
+      "loss": 0.3019,
+      "step": 1800
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 0.0002572509457755359,
+      "loss": 0.299,
+      "step": 1810
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 0.00025683060109289615,
+      "loss": 0.3062,
+      "step": 1820
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 0.00025641025641025636,
+      "loss": 0.3204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 0.0002559899117276166,
+      "loss": 0.3961,
+      "step": 1840
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 0.00025556956704497687,
+      "loss": 0.3198,
+      "step": 1850
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 0.00025514922236233707,
+      "loss": 0.3035,
+      "step": 1860
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.00025472887767969733,
+      "loss": 0.384,
+      "step": 1870
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 0.00025430853299705753,
+      "loss": 0.3199,
+      "step": 1880
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 0.0002538881883144178,
+      "loss": 0.3662,
+      "step": 1890
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.00025346784363177805,
+      "loss": 0.3327,
+      "step": 1900
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 0.00025304749894913825,
+      "loss": 0.3854,
+      "step": 1910
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 0.0002526271542664985,
+      "loss": 0.3952,
+      "step": 1920
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 0.00025220680958385876,
+      "loss": 0.3061,
+      "step": 1930
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 0.00025178646490121897,
+      "loss": 0.3574,
+      "step": 1940
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 0.0002513661202185792,
+      "loss": 0.2689,
+      "step": 1950
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 0.0002509457755359394,
+      "loss": 0.348,
+      "step": 1960
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 0.0002505254308532997,
+      "loss": 0.3279,
+      "step": 1970
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 0.00025010508617065994,
+      "loss": 0.3053,
+      "step": 1980
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 0.00024968474148802014,
+      "loss": 0.2891,
+      "step": 1990
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 0.0002492643968053804,
+      "loss": 0.3194,
+      "step": 2000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 0.00024884405212274066,
+      "loss": 0.3149,
+      "step": 2010
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 0.00024842370744010086,
+      "loss": 0.3465,
+      "step": 2020
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 0.0002480033627574611,
+      "loss": 0.2856,
+      "step": 2030
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 0.0002475830180748213,
+      "loss": 0.2747,
+      "step": 2040
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 0.0002471626733921816,
+      "loss": 0.3623,
+      "step": 2050
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.00024674232870954183,
+      "loss": 0.2506,
+      "step": 2060
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 0.00024632198402690204,
+      "loss": 0.2991,
+      "step": 2070
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 0.0002459016393442623,
+      "loss": 0.3268,
+      "step": 2080
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 0.0002454812946616225,
+      "loss": 0.3914,
+      "step": 2090
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 0.00024506094997898275,
+      "loss": 0.285,
+      "step": 2100
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 0.000244640605296343,
+      "loss": 0.3335,
+      "step": 2110
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 0.0002442202606137032,
+      "loss": 0.2386,
+      "step": 2120
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 0.00024379991593106347,
+      "loss": 0.3149,
+      "step": 2130
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 0.00024337957124842367,
+      "loss": 0.4078,
+      "step": 2140
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 0.00024295922656578393,
+      "loss": 0.2873,
+      "step": 2150
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 0.00024253888188314413,
+      "loss": 0.3443,
+      "step": 2160
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 0.0002421185372005044,
+      "loss": 0.3807,
+      "step": 2170
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 0.00024169819251786465,
+      "loss": 0.3259,
+      "step": 2180
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 0.00024127784783522485,
+      "loss": 0.2754,
+      "step": 2190
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 0.0002408575031525851,
+      "loss": 0.3341,
+      "step": 2200
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 0.0002404371584699453,
+      "loss": 0.2992,
+      "step": 2210
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 0.00024001681378730557,
+      "loss": 0.3103,
+      "step": 2220
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 0.0002395964691046658,
+      "loss": 0.3618,
+      "step": 2230
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 0.00023917612442202603,
+      "loss": 0.3182,
+      "step": 2240
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 0.00023875577973938629,
+      "loss": 0.3809,
+      "step": 2250
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 0.00023833543505674652,
+      "loss": 0.3649,
+      "step": 2260
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 0.00023791509037410675,
+      "loss": 0.3148,
+      "step": 2270
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 0.00023749474569146698,
+      "loss": 0.3599,
+      "step": 2280
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 0.00023707440100882723,
+      "loss": 0.3909,
+      "step": 2290
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 0.00023665405632618746,
+      "loss": 0.355,
+      "step": 2300
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 0.0002362337116435477,
+      "loss": 0.3217,
+      "step": 2310
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 0.00023581336696090792,
+      "loss": 0.3261,
+      "step": 2320
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 0.00023539302227826815,
+      "loss": 0.3025,
+      "step": 2330
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 0.0002349726775956284,
+      "loss": 0.3042,
+      "step": 2340
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 0.0002345523329129886,
+      "loss": 0.3965,
+      "step": 2350
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 0.00023413198823034887,
+      "loss": 0.3394,
+      "step": 2360
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 0.00023371164354770913,
+      "loss": 0.2898,
+      "step": 2370
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.3341022729873657,
+      "eval_runtime": 8152.6007,
+      "eval_samples_per_second": 3.047,
+      "eval_steps_per_second": 1.524,
+      "step": 2379
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.00023329129886506933,
+      "loss": 0.3433,
+      "step": 2380
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 0.0002328709541824296,
+      "loss": 0.3209,
+      "step": 2390
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 0.0002324506094997898,
+      "loss": 0.3302,
+      "step": 2400
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 0.00023203026481715005,
+      "loss": 0.3329,
+      "step": 2410
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 0.00023160992013451025,
+      "loss": 0.3921,
+      "step": 2420
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 0.0002311895754518705,
+      "loss": 0.3522,
+      "step": 2430
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 0.00023076923076923076,
+      "loss": 0.3266,
+      "step": 2440
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 0.00023034888608659097,
+      "loss": 0.3438,
+      "step": 2450
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 0.00022992854140395122,
+      "loss": 0.3607,
+      "step": 2460
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 0.00022950819672131145,
+      "loss": 0.2718,
+      "step": 2470
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 0.00022908785203867168,
+      "loss": 0.3434,
+      "step": 2480
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 0.00022866750735603194,
+      "loss": 0.3075,
+      "step": 2490
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 0.00022824716267339217,
+      "loss": 0.3503,
+      "step": 2500
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 0.0002278268179907524,
+      "loss": 0.2485,
+      "step": 2510
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 0.00022740647330811263,
+      "loss": 0.3542,
+      "step": 2520
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 0.00022698612862547286,
+      "loss": 0.2805,
+      "step": 2530
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.0002265657839428331,
+      "loss": 0.2834,
+      "step": 2540
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 0.00022614543926019335,
+      "loss": 0.3816,
+      "step": 2550
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 0.00022572509457755358,
+      "loss": 0.3208,
+      "step": 2560
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 0.0002253047498949138,
+      "loss": 0.4104,
+      "step": 2570
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 0.00022488440521227406,
+      "loss": 0.3087,
+      "step": 2580
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 0.00022446406052963427,
+      "loss": 0.382,
+      "step": 2590
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 0.00022404371584699452,
+      "loss": 0.3579,
+      "step": 2600
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 0.00022362337116435473,
+      "loss": 0.332,
+      "step": 2610
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 0.00022320302648171498,
+      "loss": 0.3802,
+      "step": 2620
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 0.00022278268179907524,
+      "loss": 0.3524,
+      "step": 2630
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 0.00022236233711643544,
+      "loss": 0.3372,
+      "step": 2640
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 0.0002219419924337957,
+      "loss": 0.3096,
+      "step": 2650
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 0.0002215216477511559,
+      "loss": 0.3092,
+      "step": 2660
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 0.00022110130306851616,
+      "loss": 0.3053,
+      "step": 2670
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 0.0002206809583858764,
+      "loss": 0.3133,
+      "step": 2680
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 0.00022026061370323662,
+      "loss": 0.3265,
+      "step": 2690
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.00021984026902059688,
+      "loss": 0.3675,
+      "step": 2700
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 0.0002194199243379571,
+      "loss": 0.2895,
+      "step": 2710
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 0.00021899957965531734,
+      "loss": 0.3192,
+      "step": 2720
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 0.00021857923497267757,
+      "loss": 0.2828,
+      "step": 2730
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 0.00021815889029003783,
+      "loss": 0.2881,
+      "step": 2740
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 0.00021773854560739806,
+      "loss": 0.3394,
+      "step": 2750
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 0.00021731820092475829,
+      "loss": 0.2679,
+      "step": 2760
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 0.00021689785624211852,
+      "loss": 0.3027,
+      "step": 2770
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 0.00021647751155947875,
+      "loss": 0.3233,
+      "step": 2780
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 0.000216057166876839,
+      "loss": 0.2809,
+      "step": 2790
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 0.0002156368221941992,
+      "loss": 0.3219,
+      "step": 2800
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 0.00021521647751155946,
+      "loss": 0.3301,
+      "step": 2810
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 0.00021479613282891972,
+      "loss": 0.3535,
+      "step": 2820
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 0.00021437578814627992,
+      "loss": 0.3135,
+      "step": 2830
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 0.00021395544346364018,
+      "loss": 0.31,
+      "step": 2840
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 0.00021353509878100038,
+      "loss": 0.3544,
+      "step": 2850
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 0.00021311475409836064,
+      "loss": 0.3844,
+      "step": 2860
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 0.00021269440941572087,
+      "loss": 0.2751,
+      "step": 2870
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 0.0002122740647330811,
+      "loss": 0.3444,
+      "step": 2880
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 0.00021185372005044136,
+      "loss": 0.2892,
+      "step": 2890
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 0.00021143337536780156,
+      "loss": 0.3636,
+      "step": 2900
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 0.00021101303068516182,
+      "loss": 0.3965,
+      "step": 2910
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 0.00021059268600252205,
+      "loss": 0.2829,
+      "step": 2920
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 0.00021017234131988228,
+      "loss": 0.3087,
+      "step": 2930
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 0.00020975199663724253,
+      "loss": 0.3737,
+      "step": 2940
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 0.00020933165195460276,
+      "loss": 0.368,
+      "step": 2950
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 0.000208911307271963,
+      "loss": 0.2916,
+      "step": 2960
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 0.00020849096258932322,
+      "loss": 0.3456,
+      "step": 2970
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 0.00020807061790668348,
+      "loss": 0.3588,
+      "step": 2980
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 0.00020765027322404368,
+      "loss": 0.3413,
+      "step": 2990
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 0.00020722992854140394,
+      "loss": 0.3483,
+      "step": 3000
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 0.00020680958385876417,
+      "loss": 0.3859,
+      "step": 3010
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 0.0002063892391761244,
+      "loss": 0.3477,
+      "step": 3020
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 0.00020596889449348466,
+      "loss": 0.3527,
+      "step": 3030
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 0.00020554854981084486,
+      "loss": 0.3384,
+      "step": 3040
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.00020512820512820512,
+      "loss": 0.368,
+      "step": 3050
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 0.00020470786044556532,
+      "loss": 0.2576,
+      "step": 3060
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 0.00020428751576292558,
+      "loss": 0.3343,
+      "step": 3070
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 0.00020386717108028583,
+      "loss": 0.3116,
+      "step": 3080
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 0.00020344682639764604,
+      "loss": 0.3627,
+      "step": 3090
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 0.0002030264817150063,
+      "loss": 0.3463,
+      "step": 3100
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 0.00020260613703236652,
+      "loss": 0.3301,
+      "step": 3110
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 0.00020218579234972675,
+      "loss": 0.308,
+      "step": 3120
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 0.000201765447667087,
+      "loss": 0.3252,
+      "step": 3130
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 0.00020134510298444721,
+      "loss": 0.3865,
+      "step": 3140
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 0.00020092475830180747,
+      "loss": 0.2768,
+      "step": 3150
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 0.0002005044136191677,
+      "loss": 0.3527,
+      "step": 3160
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 0.00020008406893652793,
+      "loss": 0.3126,
+      "step": 3170
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.3362921476364136,
+      "eval_runtime": 509.7321,
+      "eval_samples_per_second": 48.739,
+      "eval_steps_per_second": 24.37,
+      "step": 3173
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 0.00019966372425388816,
+      "loss": 0.3647,
+      "step": 3180
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 0.00019924337957124842,
+      "loss": 0.3627,
+      "step": 3190
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 0.00019882303488860865,
+      "loss": 0.3192,
+      "step": 3200
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 0.00019840269020596888,
+      "loss": 0.3647,
+      "step": 3210
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 0.00019798234552332914,
+      "loss": 0.3157,
+      "step": 3220
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 0.00019756200084068934,
+      "loss": 0.3907,
+      "step": 3230
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 0.0001971416561580496,
+      "loss": 0.3215,
+      "step": 3240
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 0.0001967213114754098,
+      "loss": 0.3396,
+      "step": 3250
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 0.00019630096679277006,
+      "loss": 0.3507,
+      "step": 3260
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 0.0001958806221101303,
+      "loss": 0.3467,
+      "step": 3270
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 0.00019546027742749051,
+      "loss": 0.3503,
+      "step": 3280
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 0.00019503993274485077,
+      "loss": 0.3081,
+      "step": 3290
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 0.00019461958806221097,
+      "loss": 0.308,
+      "step": 3300
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 0.00019419924337957123,
+      "loss": 0.3443,
+      "step": 3310
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 0.00019377889869693146,
+      "loss": 0.2636,
+      "step": 3320
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 0.0001933585540142917,
+      "loss": 0.3729,
+      "step": 3330
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 0.00019293820933165195,
+      "loss": 0.3245,
+      "step": 3340
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 0.00019251786464901218,
+      "loss": 0.3124,
+      "step": 3350
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 0.0001920975199663724,
+      "loss": 0.3995,
+      "step": 3360
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 0.00019167717528373264,
+      "loss": 0.3323,
+      "step": 3370
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 0.00019125683060109287,
+      "loss": 0.3126,
+      "step": 3380
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 0.00019083648591845313,
+      "loss": 0.285,
+      "step": 3390
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 0.00019041614123581336,
+      "loss": 0.3357,
+      "step": 3400
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 0.00018999579655317359,
+      "loss": 0.3878,
+      "step": 3410
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 0.00018957545187053382,
+      "loss": 0.3503,
+      "step": 3420
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 0.00018915510718789407,
+      "loss": 0.2954,
+      "step": 3430
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 0.00018873476250525428,
+      "loss": 0.3482,
+      "step": 3440
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 0.00018831441782261453,
+      "loss": 0.3512,
+      "step": 3450
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 0.0001878940731399748,
+      "loss": 0.4073,
+      "step": 3460
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 0.000187473728457335,
+      "loss": 0.3491,
+      "step": 3470
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 0.00018705338377469525,
+      "loss": 0.2629,
+      "step": 3480
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 0.00018663303909205545,
+      "loss": 0.3332,
+      "step": 3490
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 0.0001862126944094157,
+      "loss": 0.3113,
+      "step": 3500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 0.0001857923497267759,
+      "loss": 0.3737,
+      "step": 3510
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 0.00018537200504413617,
+      "loss": 0.3043,
+      "step": 3520
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 0.00018495166036149643,
+      "loss": 0.3272,
+      "step": 3530
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 0.00018453131567885663,
+      "loss": 0.3472,
+      "step": 3540
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 0.0001841109709962169,
+      "loss": 0.2957,
+      "step": 3550
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 0.00018369062631357712,
+      "loss": 0.3327,
+      "step": 3560
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 0.00018327028163093735,
+      "loss": 0.3657,
+      "step": 3570
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 0.0001828499369482976,
+      "loss": 0.2802,
+      "step": 3580
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 0.0001824295922656578,
+      "loss": 0.3196,
+      "step": 3590
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 0.00018200924758301806,
+      "loss": 0.3366,
+      "step": 3600
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 0.0001815889029003783,
+      "loss": 0.3445,
+      "step": 3610
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 0.00018116855821773852,
+      "loss": 0.2766,
+      "step": 3620
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 0.00018074821353509875,
+      "loss": 0.3672,
+      "step": 3630
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 0.000180327868852459,
+      "loss": 0.3043,
+      "step": 3640
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 0.00017990752416981924,
+      "loss": 0.399,
+      "step": 3650
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 0.00017948717948717947,
+      "loss": 0.2506,
+      "step": 3660
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 0.00017906683480453973,
+      "loss": 0.3712,
+      "step": 3670
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 0.00017864649012189993,
+      "loss": 0.2746,
+      "step": 3680
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 0.0001782261454392602,
+      "loss": 0.3563,
+      "step": 3690
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 0.0001778058007566204,
+      "loss": 0.2542,
+      "step": 3700
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 0.00017738545607398065,
+      "loss": 0.3648,
+      "step": 3710
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 0.0001769651113913409,
+      "loss": 0.3057,
+      "step": 3720
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 0.0001765447667087011,
+      "loss": 0.333,
+      "step": 3730
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 0.00017612442202606136,
+      "loss": 0.2783,
+      "step": 3740
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 0.00017570407734342157,
+      "loss": 0.2753,
+      "step": 3750
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 0.00017528373266078182,
+      "loss": 0.2699,
+      "step": 3760
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 0.00017486338797814208,
+      "loss": 0.3287,
+      "step": 3770
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 0.00017444304329550228,
+      "loss": 0.344,
+      "step": 3780
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 0.00017402269861286254,
+      "loss": 0.3526,
+      "step": 3790
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 0.00017360235393022277,
+      "loss": 0.2776,
+      "step": 3800
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 0.000173182009247583,
+      "loss": 0.3536,
+      "step": 3810
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 0.00017276166456494323,
+      "loss": 0.2467,
+      "step": 3820
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 0.00017234131988230346,
+      "loss": 0.3359,
+      "step": 3830
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 0.00017192097519966372,
+      "loss": 0.4014,
+      "step": 3840
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 0.00017150063051702395,
+      "loss": 0.3726,
+      "step": 3850
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 0.00017108028583438418,
+      "loss": 0.3849,
+      "step": 3860
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 0.0001706599411517444,
+      "loss": 0.3101,
+      "step": 3870
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 0.00017023959646910467,
+      "loss": 0.3054,
+      "step": 3880
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 0.00016981925178646487,
+      "loss": 0.2524,
+      "step": 3890
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 0.00016939890710382513,
+      "loss": 0.3549,
+      "step": 3900
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 0.00016897856242118538,
+      "loss": 0.3411,
+      "step": 3910
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 0.00016855821773854559,
+      "loss": 0.38,
+      "step": 3920
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 0.00016813787305590584,
+      "loss": 0.3008,
+      "step": 3930
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 0.00016771752837326605,
+      "loss": 0.3259,
+      "step": 3940
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 0.0001672971836906263,
+      "loss": 0.3247,
+      "step": 3950
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 0.0001668768390079865,
+      "loss": 0.3753,
+      "step": 3960
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.3308884799480438,
+      "eval_runtime": 509.4752,
+      "eval_samples_per_second": 48.764,
+      "eval_steps_per_second": 24.382,
+      "step": 3966
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 0.00016645649432534676,
+      "loss": 0.3331,
+      "step": 3970
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 0.00016603614964270702,
+      "loss": 0.3138,
+      "step": 3980
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 0.00016561580496006722,
+      "loss": 0.3933,
+      "step": 3990
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 0.00016519546027742748,
+      "loss": 0.306,
+      "step": 4000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 0.0001647751155947877,
+      "loss": 0.4105,
+      "step": 4010
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 0.00016435477091214794,
+      "loss": 0.2734,
+      "step": 4020
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 0.0001639344262295082,
+      "loss": 0.2652,
+      "step": 4030
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 0.00016351408154686843,
+      "loss": 0.3264,
+      "step": 4040
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 0.00016309373686422866,
+      "loss": 0.2907,
+      "step": 4050
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 0.0001626733921815889,
+      "loss": 0.3121,
+      "step": 4060
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 0.00016225304749894912,
+      "loss": 0.3871,
+      "step": 4070
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 0.00016183270281630935,
+      "loss": 0.297,
+      "step": 4080
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 0.0001614123581336696,
+      "loss": 0.3253,
+      "step": 4090
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 0.00016099201345102983,
+      "loss": 0.366,
+      "step": 4100
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 0.00016057166876839006,
+      "loss": 0.2644,
+      "step": 4110
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 0.00016015132408575032,
+      "loss": 0.2977,
+      "step": 4120
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 0.00015973097940311052,
+      "loss": 0.33,
+      "step": 4130
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 0.00015931063472047078,
+      "loss": 0.3278,
+      "step": 4140
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 0.00015889029003783098,
+      "loss": 0.3761,
+      "step": 4150
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 0.00015846994535519124,
+      "loss": 0.3108,
+      "step": 4160
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 0.0001580496006725515,
+      "loss": 0.3465,
+      "step": 4170
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 0.0001576292559899117,
+      "loss": 0.3401,
+      "step": 4180
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 0.00015720891130727196,
+      "loss": 0.356,
+      "step": 4190
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 0.00015678856662463216,
+      "loss": 0.3053,
+      "step": 4200
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 0.00015636822194199242,
+      "loss": 0.3047,
+      "step": 4210
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 0.00015594787725935267,
+      "loss": 0.3608,
+      "step": 4220
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 0.00015552753257671288,
+      "loss": 0.3309,
+      "step": 4230
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 0.00015510718789407313,
+      "loss": 0.3882,
+      "step": 4240
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 0.00015468684321143336,
+      "loss": 0.3731,
+      "step": 4250
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 0.0001542664985287936,
+      "loss": 0.3403,
+      "step": 4260
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 0.00015384615384615382,
+      "loss": 0.2838,
+      "step": 4270
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 0.00015342580916351408,
+      "loss": 0.2763,
+      "step": 4280
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 0.0001530054644808743,
+      "loss": 0.277,
+      "step": 4290
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 0.00015258511979823454,
+      "loss": 0.3273,
+      "step": 4300
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 0.00015216477511559477,
+      "loss": 0.3316,
+      "step": 4310
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 0.000151744430432955,
+      "loss": 0.3102,
+      "step": 4320
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 0.00015132408575031526,
+      "loss": 0.2981,
+      "step": 4330
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 0.00015090374106767546,
+      "loss": 0.3252,
+      "step": 4340
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 0.00015048339638503572,
+      "loss": 0.3686,
+      "step": 4350
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 0.00015006305170239598,
+      "loss": 0.3778,
+      "step": 4360
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 0.00014964270701975618,
+      "loss": 0.3539,
+      "step": 4370
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 0.00014922236233711644,
+      "loss": 0.3741,
+      "step": 4380
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 0.00014880201765447667,
+      "loss": 0.3339,
+      "step": 4390
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 0.0001483816729718369,
+      "loss": 0.3037,
+      "step": 4400
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 0.00014796132828919713,
+      "loss": 0.3402,
+      "step": 4410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 0.00014754098360655736,
+      "loss": 0.2921,
+      "step": 4420
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 0.00014712063892391759,
+      "loss": 0.294,
+      "step": 4430
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 0.00014670029424127782,
+      "loss": 0.3263,
+      "step": 4440
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 0.00014627994955863807,
+      "loss": 0.3616,
+      "step": 4450
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 0.0001458596048759983,
+      "loss": 0.4128,
+      "step": 4460
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 0.00014543926019335853,
+      "loss": 0.346,
+      "step": 4470
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 0.00014501891551071876,
+      "loss": 0.3338,
+      "step": 4480
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 0.00014459857082807902,
+      "loss": 0.3784,
+      "step": 4490
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 0.00014417822614543925,
+      "loss": 0.2997,
+      "step": 4500
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 0.00014375788146279948,
+      "loss": 0.4233,
+      "step": 4510
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 0.00014333753678015974,
+      "loss": 0.3164,
+      "step": 4520
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 0.00014291719209751997,
+      "loss": 0.3837,
+      "step": 4530
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 0.0001424968474148802,
+      "loss": 0.2968,
+      "step": 4540
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 0.00014207650273224043,
+      "loss": 0.2592,
+      "step": 4550
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 0.00014165615804960066,
+      "loss": 0.3261,
+      "step": 4560
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 0.0001412358133669609,
+      "loss": 0.3475,
+      "step": 4570
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 0.00014081546868432114,
+      "loss": 0.3477,
+      "step": 4580
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 0.00014039512400168137,
+      "loss": 0.29,
+      "step": 4590
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 0.0001399747793190416,
+      "loss": 0.3797,
+      "step": 4600
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 0.00013955443463640183,
+      "loss": 0.2906,
+      "step": 4610
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 0.00013913408995376206,
+      "loss": 0.3484,
+      "step": 4620
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 0.0001387137452711223,
+      "loss": 0.3059,
+      "step": 4630
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 0.00013829340058848255,
+      "loss": 0.3363,
+      "step": 4640
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 0.00013787305590584278,
+      "loss": 0.356,
+      "step": 4650
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 0.000137452711223203,
+      "loss": 0.2837,
+      "step": 4660
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 0.00013703236654056324,
+      "loss": 0.3176,
+      "step": 4670
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 0.00013661202185792347,
+      "loss": 0.2976,
+      "step": 4680
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 0.00013619167717528373,
+      "loss": 0.2907,
+      "step": 4690
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 0.00013577133249264396,
+      "loss": 0.3671,
+      "step": 4700
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 0.0001353509878100042,
+      "loss": 0.3118,
+      "step": 4710
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 0.00013493064312736442,
+      "loss": 0.3113,
+      "step": 4720
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 0.00013451029844472467,
+      "loss": 0.319,
+      "step": 4730
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 0.0001340899537620849,
+      "loss": 0.3043,
+      "step": 4740
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 0.00013366960907944513,
+      "loss": 0.3617,
+      "step": 4750
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.33245739340782166,
+      "eval_runtime": 509.9573,
+      "eval_samples_per_second": 48.718,
+      "eval_steps_per_second": 24.359,
+      "step": 4759
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 0.00013324926439680536,
+      "loss": 0.3967,
+      "step": 4760
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 0.00013282891971416562,
+      "loss": 0.3046,
+      "step": 4770
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 0.00013240857503152585,
+      "loss": 0.303,
+      "step": 4780
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 0.00013198823034888608,
+      "loss": 0.2971,
+      "step": 4790
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 0.0001315678856662463,
+      "loss": 0.3187,
+      "step": 4800
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 0.00013114754098360654,
+      "loss": 0.3411,
+      "step": 4810
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 0.00013072719630096677,
+      "loss": 0.333,
+      "step": 4820
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 0.00013030685161832703,
+      "loss": 0.3181,
+      "step": 4830
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 0.00012988650693568726,
+      "loss": 0.3046,
+      "step": 4840
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 0.0001294661622530475,
+      "loss": 0.3539,
+      "step": 4850
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 0.00012904581757040772,
+      "loss": 0.2974,
+      "step": 4860
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 0.00012862547288776795,
+      "loss": 0.3189,
+      "step": 4870
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 0.00012820512820512818,
+      "loss": 0.3038,
+      "step": 4880
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 0.00012778478352248843,
+      "loss": 0.3741,
+      "step": 4890
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 0.00012736443883984866,
+      "loss": 0.3264,
+      "step": 4900
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 0.0001269440941572089,
+      "loss": 0.2311,
+      "step": 4910
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 0.00012652374947456912,
+      "loss": 0.3001,
+      "step": 4920
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 0.00012610340479192938,
+      "loss": 0.3067,
+      "step": 4930
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 0.0001256830601092896,
+      "loss": 0.3493,
+      "step": 4940
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 0.00012526271542664984,
+      "loss": 0.2825,
+      "step": 4950
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 0.00012484237074401007,
+      "loss": 0.4099,
+      "step": 4960
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 0.00012442202606137033,
+      "loss": 0.2912,
+      "step": 4970
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 0.00012400168137873056,
+      "loss": 0.3586,
+      "step": 4980
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 0.0001235813366960908,
+      "loss": 0.2966,
+      "step": 4990
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 0.00012316099201345102,
+      "loss": 0.306,
+      "step": 5000
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 0.00012274064733081125,
+      "loss": 0.4236,
+      "step": 5010
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 0.0001223203026481715,
+      "loss": 0.3258,
+      "step": 5020
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 0.00012189995796553174,
+      "loss": 0.3197,
+      "step": 5030
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 0.00012147961328289197,
+      "loss": 0.2154,
+      "step": 5040
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 0.0001210592686002522,
+      "loss": 0.3038,
+      "step": 5050
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 0.00012063892391761243,
+      "loss": 0.4093,
+      "step": 5060
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 0.00012021857923497266,
+      "loss": 0.3657,
+      "step": 5070
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 0.0001197982345523329,
+      "loss": 0.325,
+      "step": 5080
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 0.00011937788986969314,
+      "loss": 0.3583,
+      "step": 5090
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 0.00011895754518705337,
+      "loss": 0.3518,
+      "step": 5100
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 0.00011853720050441362,
+      "loss": 0.3191,
+      "step": 5110
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 0.00011811685582177385,
+      "loss": 0.3688,
+      "step": 5120
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 0.00011769651113913408,
+      "loss": 0.3046,
+      "step": 5130
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 0.0001172761664564943,
+      "loss": 0.2689,
+      "step": 5140
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 0.00011685582177385456,
+      "loss": 0.2747,
+      "step": 5150
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 0.0001164354770912148,
+      "loss": 0.4487,
+      "step": 5160
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 0.00011601513240857502,
+      "loss": 0.3329,
+      "step": 5170
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 0.00011559478772593525,
+      "loss": 0.2969,
+      "step": 5180
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 0.00011517444304329548,
+      "loss": 0.3122,
+      "step": 5190
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 0.00011475409836065573,
+      "loss": 0.3464,
+      "step": 5200
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 0.00011433375367801597,
+      "loss": 0.3466,
+      "step": 5210
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 0.0001139134089953762,
+      "loss": 0.3353,
+      "step": 5220
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 0.00011349306431273643,
+      "loss": 0.3462,
+      "step": 5230
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 0.00011307271963009667,
+      "loss": 0.3692,
+      "step": 5240
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 0.0001126523749474569,
+      "loss": 0.2234,
+      "step": 5250
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 0.00011223203026481713,
+      "loss": 0.2973,
+      "step": 5260
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 0.00011181168558217736,
+      "loss": 0.2808,
+      "step": 5270
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 0.00011139134089953762,
+      "loss": 0.4014,
+      "step": 5280
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 0.00011097099621689785,
+      "loss": 0.3689,
+      "step": 5290
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 0.00011055065153425808,
+      "loss": 0.3302,
+      "step": 5300
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 0.00011013030685161831,
+      "loss": 0.3392,
+      "step": 5310
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 0.00010970996216897855,
+      "loss": 0.3331,
+      "step": 5320
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 0.00010928961748633878,
+      "loss": 0.3928,
+      "step": 5330
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 0.00010886927280369903,
+      "loss": 0.3647,
+      "step": 5340
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 0.00010844892812105926,
+      "loss": 0.2881,
+      "step": 5350
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 0.0001080285834384195,
+      "loss": 0.3195,
+      "step": 5360
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 0.00010760823875577973,
+      "loss": 0.3061,
+      "step": 5370
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 0.00010718789407313996,
+      "loss": 0.3044,
+      "step": 5380
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 0.00010676754939050019,
+      "loss": 0.3549,
+      "step": 5390
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 0.00010634720470786043,
+      "loss": 0.4058,
+      "step": 5400
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 0.00010592686002522068,
+      "loss": 0.3338,
+      "step": 5410
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 0.00010550651534258091,
+      "loss": 0.2252,
+      "step": 5420
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 0.00010508617065994114,
+      "loss": 0.3422,
+      "step": 5430
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 0.00010466582597730138,
+      "loss": 0.3434,
+      "step": 5440
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 0.00010424548129466161,
+      "loss": 0.3413,
+      "step": 5450
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 0.00010382513661202184,
+      "loss": 0.3875,
+      "step": 5460
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 0.00010340479192938209,
+      "loss": 0.4227,
+      "step": 5470
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 0.00010298444724674233,
+      "loss": 0.308,
+      "step": 5480
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 0.00010256410256410256,
+      "loss": 0.3449,
+      "step": 5490
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 0.00010214375788146279,
+      "loss": 0.334,
+      "step": 5500
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 0.00010172341319882302,
+      "loss": 0.3527,
+      "step": 5510
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 0.00010130306851618326,
+      "loss": 0.2319,
+      "step": 5520
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 0.0001008827238335435,
+      "loss": 0.3336,
+      "step": 5530
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 0.00010046237915090374,
+      "loss": 0.4377,
+      "step": 5540
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 0.00010004203446826397,
+      "loss": 0.3453,
+      "step": 5550
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.33148393034935,
+      "eval_runtime": 509.701,
+      "eval_samples_per_second": 48.742,
+      "eval_steps_per_second": 24.371,
+      "step": 5552
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 9.962168978562421e-05,
+      "loss": 0.3489,
+      "step": 5560
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 9.920134510298444e-05,
+      "loss": 0.3478,
+      "step": 5570
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 9.878100042034467e-05,
+      "loss": 0.34,
+      "step": 5580
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 9.83606557377049e-05,
+      "loss": 0.2938,
+      "step": 5590
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 9.794031105506516e-05,
+      "loss": 0.3057,
+      "step": 5600
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 9.751996637242539e-05,
+      "loss": 0.3185,
+      "step": 5610
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 9.709962168978562e-05,
+      "loss": 0.3963,
+      "step": 5620
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 9.667927700714585e-05,
+      "loss": 0.3252,
+      "step": 5630
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 9.625893232450609e-05,
+      "loss": 0.3181,
+      "step": 5640
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 9.583858764186632e-05,
+      "loss": 0.3184,
+      "step": 5650
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 9.541824295922656e-05,
+      "loss": 0.3113,
+      "step": 5660
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 9.499789827658679e-05,
+      "loss": 0.2968,
+      "step": 5670
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 9.457755359394704e-05,
+      "loss": 0.3257,
+      "step": 5680
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 9.415720891130727e-05,
+      "loss": 0.2974,
+      "step": 5690
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 9.37368642286675e-05,
+      "loss": 0.2971,
+      "step": 5700
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 9.331651954602773e-05,
+      "loss": 0.29,
+      "step": 5710
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 9.289617486338796e-05,
+      "loss": 0.3821,
+      "step": 5720
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 9.247583018074821e-05,
+      "loss": 0.3632,
+      "step": 5730
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 9.205548549810844e-05,
+      "loss": 0.3589,
+      "step": 5740
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 9.163514081546867e-05,
+      "loss": 0.2663,
+      "step": 5750
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 9.12147961328289e-05,
+      "loss": 0.3121,
+      "step": 5760
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 9.079445145018915e-05,
+      "loss": 0.3527,
+      "step": 5770
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 9.037410676754938e-05,
+      "loss": 0.3045,
+      "step": 5780
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 8.995376208490962e-05,
+      "loss": 0.3109,
+      "step": 5790
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 8.953341740226986e-05,
+      "loss": 0.3256,
+      "step": 5800
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 8.91130727196301e-05,
+      "loss": 0.411,
+      "step": 5810
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 8.869272803699032e-05,
+      "loss": 0.3603,
+      "step": 5820
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 8.827238335435055e-05,
+      "loss": 0.4049,
+      "step": 5830
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 8.785203867171078e-05,
+      "loss": 0.3386,
+      "step": 5840
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 8.743169398907104e-05,
+      "loss": 0.2656,
+      "step": 5850
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 8.701134930643127e-05,
+      "loss": 0.2769,
+      "step": 5860
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 8.65910046237915e-05,
+      "loss": 0.3055,
+      "step": 5870
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 8.617065994115173e-05,
+      "loss": 0.341,
+      "step": 5880
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 8.575031525851197e-05,
+      "loss": 0.3116,
+      "step": 5890
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 8.53299705758722e-05,
+      "loss": 0.3873,
+      "step": 5900
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 8.490962589323243e-05,
+      "loss": 0.2296,
+      "step": 5910
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 8.448928121059269e-05,
+      "loss": 0.275,
+      "step": 5920
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 8.406893652795292e-05,
+      "loss": 0.3482,
+      "step": 5930
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 8.364859184531315e-05,
+      "loss": 0.3619,
+      "step": 5940
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 8.322824716267338e-05,
+      "loss": 0.3533,
+      "step": 5950
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 8.280790248003361e-05,
+      "loss": 0.3397,
+      "step": 5960
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 8.238755779739385e-05,
+      "loss": 0.318,
+      "step": 5970
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 8.19672131147541e-05,
+      "loss": 0.3619,
+      "step": 5980
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 8.154686843211433e-05,
+      "loss": 0.3263,
+      "step": 5990
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 8.112652374947456e-05,
+      "loss": 0.3397,
+      "step": 6000
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 8.07061790668348e-05,
+      "loss": 0.3253,
+      "step": 6010
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 8.028583438419503e-05,
+      "loss": 0.339,
+      "step": 6020
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 7.986548970155526e-05,
+      "loss": 0.3603,
+      "step": 6030
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 7.944514501891549e-05,
+      "loss": 0.3261,
+      "step": 6040
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 7.902480033627575e-05,
+      "loss": 0.3186,
+      "step": 6050
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 7.860445565363598e-05,
+      "loss": 0.3682,
+      "step": 6060
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 7.818411097099621e-05,
+      "loss": 0.3794,
+      "step": 6070
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 7.776376628835644e-05,
+      "loss": 0.3129,
+      "step": 6080
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 7.734342160571668e-05,
+      "loss": 0.4197,
+      "step": 6090
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 7.692307692307691e-05,
+      "loss": 0.298,
+      "step": 6100
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 7.650273224043716e-05,
+      "loss": 0.3475,
+      "step": 6110
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 7.608238755779739e-05,
+      "loss": 0.3186,
+      "step": 6120
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 7.566204287515763e-05,
+      "loss": 0.4033,
+      "step": 6130
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 7.524169819251786e-05,
+      "loss": 0.3802,
+      "step": 6140
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 7.482135350987809e-05,
+      "loss": 0.3604,
+      "step": 6150
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 7.440100882723833e-05,
+      "loss": 0.2891,
+      "step": 6160
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 7.398066414459856e-05,
+      "loss": 0.3447,
+      "step": 6170
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 7.356031946195879e-05,
+      "loss": 0.3181,
+      "step": 6180
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 7.313997477931904e-05,
+      "loss": 0.3317,
+      "step": 6190
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 7.271963009667927e-05,
+      "loss": 0.2902,
+      "step": 6200
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 7.229928541403951e-05,
+      "loss": 0.2958,
+      "step": 6210
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 7.187894073139974e-05,
+      "loss": 0.2743,
+      "step": 6220
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 7.145859604875998e-05,
+      "loss": 0.2628,
+      "step": 6230
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 7.103825136612021e-05,
+      "loss": 0.334,
+      "step": 6240
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 7.061790668348046e-05,
+      "loss": 0.2839,
+      "step": 6250
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 7.019756200084069e-05,
+      "loss": 0.2978,
+      "step": 6260
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 6.977721731820092e-05,
+      "loss": 0.3751,
+      "step": 6270
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 6.935687263556115e-05,
+      "loss": 0.332,
+      "step": 6280
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 6.893652795292139e-05,
+      "loss": 0.3452,
+      "step": 6290
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 6.851618327028162e-05,
+      "loss": 0.3821,
+      "step": 6300
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 6.809583858764186e-05,
+      "loss": 0.3459,
+      "step": 6310
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 6.76754939050021e-05,
+      "loss": 0.3526,
+      "step": 6320
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 6.725514922236234e-05,
+      "loss": 0.296,
+      "step": 6330
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 6.683480453972257e-05,
+      "loss": 0.3337,
+      "step": 6340
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.3363693356513977,
+      "eval_runtime": 510.1332,
+      "eval_samples_per_second": 48.701,
+      "eval_steps_per_second": 24.351,
+      "step": 6346
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 6.641445985708281e-05,
+      "loss": 0.2746,
+      "step": 6350
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 6.599411517444304e-05,
+      "loss": 0.3126,
+      "step": 6360
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 6.557377049180327e-05,
+      "loss": 0.2821,
+      "step": 6370
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 6.515342580916351e-05,
+      "loss": 0.4024,
+      "step": 6380
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 6.473308112652374e-05,
+      "loss": 0.3195,
+      "step": 6390
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 6.431273644388397e-05,
+      "loss": 0.4313,
+      "step": 6400
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 6.389239176124422e-05,
+      "loss": 0.3524,
+      "step": 6410
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 6.347204707860445e-05,
+      "loss": 0.4252,
+      "step": 6420
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 6.305170239596469e-05,
+      "loss": 0.3575,
+      "step": 6430
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 6.263135771332492e-05,
+      "loss": 0.3216,
+      "step": 6440
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 6.221101303068516e-05,
+      "loss": 0.346,
+      "step": 6450
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 6.17906683480454e-05,
+      "loss": 0.3773,
+      "step": 6460
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 6.137032366540562e-05,
+      "loss": 0.3073,
+      "step": 6470
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 6.094997898276587e-05,
+      "loss": 0.3452,
+      "step": 6480
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 6.05296343001261e-05,
+      "loss": 0.3835,
+      "step": 6490
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 6.010928961748633e-05,
+      "loss": 0.3656,
+      "step": 6500
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 5.968894493484657e-05,
+      "loss": 0.3449,
+      "step": 6510
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 5.926860025220681e-05,
+      "loss": 0.319,
+      "step": 6520
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 5.884825556956704e-05,
+      "loss": 0.3729,
+      "step": 6530
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 5.842791088692728e-05,
+      "loss": 0.3595,
+      "step": 6540
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 5.800756620428751e-05,
+      "loss": 0.3593,
+      "step": 6550
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 5.758722152164774e-05,
+      "loss": 0.3522,
+      "step": 6560
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 5.7166876839007985e-05,
+      "loss": 0.3656,
+      "step": 6570
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 5.6746532156368215e-05,
+      "loss": 0.2843,
+      "step": 6580
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 5.632618747372845e-05,
+      "loss": 0.3527,
+      "step": 6590
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 5.590584279108868e-05,
+      "loss": 0.3107,
+      "step": 6600
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 5.5485498108448925e-05,
+      "loss": 0.2899,
+      "step": 6610
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 5.5065153425809155e-05,
+      "loss": 0.3113,
+      "step": 6620
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 5.464480874316939e-05,
+      "loss": 0.3764,
+      "step": 6630
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 5.422446406052963e-05,
+      "loss": 0.2625,
+      "step": 6640
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 5.3804119377889866e-05,
+      "loss": 0.3396,
+      "step": 6650
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 5.3383774695250096e-05,
+      "loss": 0.3394,
+      "step": 6660
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 5.296343001261034e-05,
+      "loss": 0.3045,
+      "step": 6670
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 5.254308532997057e-05,
+      "loss": 0.3531,
+      "step": 6680
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 5.2122740647330806e-05,
+      "loss": 0.2723,
+      "step": 6690
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 5.170239596469104e-05,
+      "loss": 0.3677,
+      "step": 6700
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 5.128205128205128e-05,
+      "loss": 0.3191,
+      "step": 6710
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 5.086170659941151e-05,
+      "loss": 0.3533,
+      "step": 6720
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 5.044136191677175e-05,
+      "loss": 0.2686,
+      "step": 6730
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 5.002101723413198e-05,
+      "loss": 0.275,
+      "step": 6740
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 4.960067255149222e-05,
+      "loss": 0.3405,
+      "step": 6750
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 4.918032786885245e-05,
+      "loss": 0.3112,
+      "step": 6760
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 4.875998318621269e-05,
+      "loss": 0.4261,
+      "step": 6770
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 4.833963850357292e-05,
+      "loss": 0.2757,
+      "step": 6780
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 4.791929382093316e-05,
+      "loss": 0.2967,
+      "step": 6790
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 4.7498949138293396e-05,
+      "loss": 0.3255,
+      "step": 6800
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 4.707860445565363e-05,
+      "loss": 0.4043,
+      "step": 6810
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 4.665825977301386e-05,
+      "loss": 0.3614,
+      "step": 6820
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 4.623791509037411e-05,
+      "loss": 0.3046,
+      "step": 6830
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 4.581757040773434e-05,
+      "loss": 0.3532,
+      "step": 6840
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 4.5397225725094573e-05,
+      "loss": 0.3114,
+      "step": 6850
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 4.497688104245481e-05,
+      "loss": 0.2897,
+      "step": 6860
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 4.455653635981505e-05,
+      "loss": 0.2971,
+      "step": 6870
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 4.413619167717528e-05,
+      "loss": 0.2696,
+      "step": 6880
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 4.371584699453552e-05,
+      "loss": 0.3037,
+      "step": 6890
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 4.329550231189575e-05,
+      "loss": 0.3404,
+      "step": 6900
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 4.287515762925599e-05,
+      "loss": 0.2828,
+      "step": 6910
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 4.245481294661622e-05,
+      "loss": 0.3464,
+      "step": 6920
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 4.203446826397646e-05,
+      "loss": 0.2829,
+      "step": 6930
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 4.161412358133669e-05,
+      "loss": 0.3396,
+      "step": 6940
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 4.119377889869693e-05,
+      "loss": 0.3116,
+      "step": 6950
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 4.0773434216057164e-05,
+      "loss": 0.3534,
+      "step": 6960
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 4.03530895334174e-05,
+      "loss": 0.3177,
+      "step": 6970
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 3.993274485077763e-05,
+      "loss": 0.3183,
+      "step": 6980
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 3.9512400168137874e-05,
+      "loss": 0.297,
+      "step": 6990
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9092055485498104e-05,
+      "loss": 0.2697,
+      "step": 7000
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 3.867171080285834e-05,
+      "loss": 0.3032,
+      "step": 7010
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 3.825136612021858e-05,
+      "loss": 0.3244,
+      "step": 7020
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.7831021437578815e-05,
+      "loss": 0.3245,
+      "step": 7030
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 3.7410676754939045e-05,
+      "loss": 0.3741,
+      "step": 7040
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 3.699033207229928e-05,
+      "loss": 0.2971,
+      "step": 7050
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.656998738965952e-05,
+      "loss": 0.2829,
+      "step": 7060
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 3.6149642707019755e-05,
+      "loss": 0.3683,
+      "step": 7070
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 3.572929802437999e-05,
+      "loss": 0.3487,
+      "step": 7080
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.530895334174023e-05,
+      "loss": 0.3199,
+      "step": 7090
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 3.488860865910046e-05,
+      "loss": 0.2771,
+      "step": 7100
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 3.4468263976460695e-05,
+      "loss": 0.3527,
+      "step": 7110
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.404791929382093e-05,
+      "loss": 0.2974,
+      "step": 7120
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 3.362757461118117e-05,
+      "loss": 0.2829,
+      "step": 7130
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.332711786031723,
+      "eval_runtime": 509.8549,
+      "eval_samples_per_second": 48.728,
+      "eval_steps_per_second": 24.364,
+      "step": 7139
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 3.3207229928541405e-05,
+      "loss": 0.3534,
+      "step": 7140
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 3.2786885245901635e-05,
+      "loss": 0.3682,
+      "step": 7150
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 3.236654056326187e-05,
+      "loss": 0.3243,
+      "step": 7160
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 3.194619588062211e-05,
+      "loss": 0.3626,
+      "step": 7170
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 3.1525851197982345e-05,
+      "loss": 0.3114,
+      "step": 7180
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.110550651534258e-05,
+      "loss": 0.2905,
+      "step": 7190
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 3.068516183270281e-05,
+      "loss": 0.3038,
+      "step": 7200
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 3.026481715006305e-05,
+      "loss": 0.392,
+      "step": 7210
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 2.9844472467423286e-05,
+      "loss": 0.3543,
+      "step": 7220
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 2.942412778478352e-05,
+      "loss": 0.3937,
+      "step": 7230
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 2.9003783102143756e-05,
+      "loss": 0.2724,
+      "step": 7240
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 2.8583438419503993e-05,
+      "loss": 0.326,
+      "step": 7250
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 2.8163093736864226e-05,
+      "loss": 0.3197,
+      "step": 7260
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 2.7742749054224463e-05,
+      "loss": 0.2846,
+      "step": 7270
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 2.7322404371584696e-05,
+      "loss": 0.3329,
+      "step": 7280
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 2.6902059688944933e-05,
+      "loss": 0.3542,
+      "step": 7290
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 2.648171500630517e-05,
+      "loss": 0.2974,
+      "step": 7300
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 2.6061370323665403e-05,
+      "loss": 0.2765,
+      "step": 7310
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 2.564102564102564e-05,
+      "loss": 0.3265,
+      "step": 7320
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 2.5220680958385876e-05,
+      "loss": 0.2828,
+      "step": 7330
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 2.480033627574611e-05,
+      "loss": 0.2894,
+      "step": 7340
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 2.4379991593106347e-05,
+      "loss": 0.3318,
+      "step": 7350
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 2.395964691046658e-05,
+      "loss": 0.3119,
+      "step": 7360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 2.3539302227826817e-05,
+      "loss": 0.2827,
+      "step": 7370
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 2.3118957545187053e-05,
+      "loss": 0.3331,
+      "step": 7380
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 2.2698612862547287e-05,
+      "loss": 0.3107,
+      "step": 7390
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 2.2278268179907523e-05,
+      "loss": 0.2939,
+      "step": 7400
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 2.185792349726776e-05,
+      "loss": 0.3974,
+      "step": 7410
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 2.1437578814627994e-05,
+      "loss": 0.2909,
+      "step": 7420
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 2.101723413198823e-05,
+      "loss": 0.268,
+      "step": 7430
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 2.0596889449348464e-05,
+      "loss": 0.3889,
+      "step": 7440
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 2.01765447667087e-05,
+      "loss": 0.3187,
+      "step": 7450
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.9756200084068937e-05,
+      "loss": 0.3466,
+      "step": 7460
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.933585540142917e-05,
+      "loss": 0.3672,
+      "step": 7470
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.8915510718789407e-05,
+      "loss": 0.3113,
+      "step": 7480
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.849516603614964e-05,
+      "loss": 0.2826,
+      "step": 7490
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.8074821353509877e-05,
+      "loss": 0.2822,
+      "step": 7500
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.7654476670870114e-05,
+      "loss": 0.3111,
+      "step": 7510
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.7234131988230348e-05,
+      "loss": 0.2607,
+      "step": 7520
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.6813787305590584e-05,
+      "loss": 0.3614,
+      "step": 7530
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.6393442622950818e-05,
+      "loss": 0.3901,
+      "step": 7540
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.5973097940311054e-05,
+      "loss": 0.4236,
+      "step": 7550
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.555275325767129e-05,
+      "loss": 0.3317,
+      "step": 7560
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.5132408575031524e-05,
+      "loss": 0.3361,
+      "step": 7570
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.471206389239176e-05,
+      "loss": 0.2569,
+      "step": 7580
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 1.4291719209751996e-05,
+      "loss": 0.3112,
+      "step": 7590
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 1.3871374527112231e-05,
+      "loss": 0.3608,
+      "step": 7600
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 1.3451029844472466e-05,
+      "loss": 0.3478,
+      "step": 7610
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 1.3030685161832701e-05,
+      "loss": 0.3728,
+      "step": 7620
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 1.2610340479192938e-05,
+      "loss": 0.2908,
+      "step": 7630
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 1.2189995796553173e-05,
+      "loss": 0.4091,
+      "step": 7640
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 1.1769651113913408e-05,
+      "loss": 0.3986,
+      "step": 7650
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 1.1349306431273643e-05,
+      "loss": 0.3454,
+      "step": 7660
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 1.092896174863388e-05,
+      "loss": 0.3265,
+      "step": 7670
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.0508617065994115e-05,
+      "loss": 0.3382,
+      "step": 7680
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 1.008827238335435e-05,
+      "loss": 0.3119,
+      "step": 7690
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 9.667927700714585e-06,
+      "loss": 0.2555,
+      "step": 7700
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 9.24758301807482e-06,
+      "loss": 0.3322,
+      "step": 7710
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 8.827238335435057e-06,
+      "loss": 0.3408,
+      "step": 7720
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 8.406893652795292e-06,
+      "loss": 0.3037,
+      "step": 7730
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 7.986548970155527e-06,
+      "loss": 0.3184,
+      "step": 7740
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 7.566204287515762e-06,
+      "loss": 0.3673,
+      "step": 7750
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 7.145859604875998e-06,
+      "loss": 0.3821,
+      "step": 7760
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 6.725514922236233e-06,
+      "loss": 0.4443,
+      "step": 7770
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 6.305170239596469e-06,
+      "loss": 0.3253,
+      "step": 7780
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 5.884825556956704e-06,
+      "loss": 0.3526,
+      "step": 7790
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 5.46448087431694e-06,
+      "loss": 0.36,
+      "step": 7800
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 5.044136191677175e-06,
+      "loss": 0.3664,
+      "step": 7810
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 4.62379150903741e-06,
+      "loss": 0.3597,
+      "step": 7820
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 4.203446826397646e-06,
+      "loss": 0.3248,
+      "step": 7830
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.783102143757881e-06,
+      "loss": 0.283,
+      "step": 7840
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 3.3627574611181166e-06,
+      "loss": 0.3326,
+      "step": 7850
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 2.942412778478352e-06,
+      "loss": 0.2613,
+      "step": 7860
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 2.5220680958385876e-06,
+      "loss": 0.2902,
+      "step": 7870
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 2.101723413198823e-06,
+      "loss": 0.3327,
+      "step": 7880
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.6813787305590583e-06,
+      "loss": 0.2974,
+      "step": 7890
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 1.2610340479192938e-06,
+      "loss": 0.3177,
+      "step": 7900
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 8.406893652795291e-07,
+      "loss": 0.3118,
+      "step": 7910
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.203446826397646e-07,
+      "loss": 0.3667,
+      "step": 7920
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 0.0,
+      "loss": 0.3189,
+      "step": 7930
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.8974400257607471,
+      "eval_loss": 0.33213362097740173,
+      "eval_runtime": 510.0772,
+      "eval_samples_per_second": 48.706,
+      "eval_steps_per_second": 24.353,
+      "step": 7930
+    },
+    {
+      "epoch": 10.0,
+      "step": 7930,
+      "total_flos": 1.3229335126002113e+19,
+      "train_loss": 0.3268948573686476,
+      "train_runtime": 17006.2717,
+      "train_samples_per_second": 14.924,
+      "train_steps_per_second": 0.466
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7930,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "total_flos": 1.3229335126002113e+19,
+  "trial_name": null,
+  "trial_params": null
+}