End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 9.99,
-    "eval_accuracy": 0.8644859813084113,
-    "eval_loss": 0.4172683358192444,
-    "eval_runtime": 69.3033,
-    "eval_samples_per_second": 61.758,
-    "eval_steps_per_second": 1.934,
-    "total_flos": 1.3257453564799912e+19,
-    "train_loss": 0.3000902961967583,
-    "train_runtime": 6004.4248,
-    "train_samples_per_second": 28.507,
-    "train_steps_per_second": 0.222
+    "epoch": 10.0,
+    "eval_accuracy": 0.8761682242990654,
+    "eval_loss": 0.38753223419189453,
+    "eval_runtime": 92.7247,
+    "eval_samples_per_second": 46.158,
+    "eval_steps_per_second": 5.77,
+    "total_flos": 1.3264660513609667e+19,
+    "train_loss": 0.36025063641717503,
+    "train_runtime": 7147.2488,
+    "train_samples_per_second": 23.949,
+    "train_steps_per_second": 0.749
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.99,
-    "eval_accuracy": 0.8644859813084113,
-    "eval_loss": 0.4172683358192444,
-    "eval_runtime": 69.3033,
-    "eval_samples_per_second": 61.758,
-    "eval_steps_per_second": 1.934
+    "epoch": 10.0,
+    "eval_accuracy": 0.8761682242990654,
+    "eval_loss": 0.38753223419189453,
+    "eval_runtime": 92.7247,
+    "eval_samples_per_second": 46.158,
+    "eval_steps_per_second": 5.77
 }

runs/Oct21_08-44-12_fd488c9ba56a/events.out.tfevents.1666349321.fd488c9ba56a.17.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90e63c3e5c01b29b3229ad22ba8f5dbf00a58c6ddcad2384170ef90d6105c5f4
+size 363

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.99,
-    "total_flos": 1.3257453564799912e+19,
-    "train_loss": 0.3000902961967583,
-    "train_runtime": 6004.4248,
-    "train_samples_per_second": 28.507,
-    "train_steps_per_second": 0.222
+    "epoch": 10.0,
+    "total_flos": 1.3264660513609667e+19,
+    "train_loss": 0.36025063641717503,
+    "train_runtime": 7147.2488,
+    "train_samples_per_second": 23.949,
+    "train_steps_per_second": 0.749
 }

trainer_state.json

@@ -1,913 +1,3325 @@
 {
-  "best_metric": 0.8644859813084113,
-  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-cassava/checkpoint-399",
-  "epoch": 9.994392523364485,
-  "global_step": 1330,
+  "best_metric": 0.8761682242990654,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-cassava/checkpoint-4280",
+  "epoch": 10.0,
+  "global_step": 5350,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.07,
-      "learning_rate": 3.7593984962406014e-06,
-      "loss": 1.6081,
+      "epoch": 0.02,
+      "learning_rate": 9.345794392523364e-07,
+      "loss": 1.6171,
       "step": 10
     },
     {
-      "epoch": 0.15,
-      "learning_rate": 7.518796992481203e-06,
-      "loss": 1.5302,
+      "epoch": 0.04,
+      "learning_rate": 1.8691588785046728e-06,
+      "loss": 1.6057,
       "step": 20
     },
     {
-      "epoch": 0.22,
-      "learning_rate": 1.1278195488721805e-05,
-      "loss": 1.3659,
+      "epoch": 0.06,
+      "learning_rate": 2.8037383177570094e-06,
+      "loss": 1.5864,
       "step": 30
     },
     {
-      "epoch": 0.3,
-      "learning_rate": 1.5037593984962406e-05,
-      "loss": 1.2237,
+      "epoch": 0.07,
+      "learning_rate": 3.7383177570093455e-06,
+      "loss": 1.5484,
       "step": 40
     },
     {
-      "epoch": 0.37,
-      "learning_rate": 1.8796992481203007e-05,
-      "loss": 1.0832,
+      "epoch": 0.09,
+      "learning_rate": 4.6728971962616825e-06,
+      "loss": 1.4915,
       "step": 50
     },
     {
-      "epoch": 0.45,
-      "learning_rate": 2.255639097744361e-05,
-      "loss": 1.02,
+      "epoch": 0.11,
+      "learning_rate": 5.607476635514019e-06,
+      "loss": 1.4295,
       "step": 60
     },
     {
-      "epoch": 0.52,
-      "learning_rate": 2.6315789473684212e-05,
-      "loss": 0.8901,
+      "epoch": 0.13,
+      "learning_rate": 6.542056074766355e-06,
+      "loss": 1.306,
       "step": 70
     },
     {
-      "epoch": 0.6,
-      "learning_rate": 3.007518796992481e-05,
-      "loss": 0.8331,
+      "epoch": 0.15,
+      "learning_rate": 7.476635514018691e-06,
+      "loss": 1.2262,
       "step": 80
     },
     {
-      "epoch": 0.67,
-      "learning_rate": 3.3834586466165414e-05,
-      "loss": 0.751,
+      "epoch": 0.17,
+      "learning_rate": 8.411214953271028e-06,
+      "loss": 1.1393,
       "step": 90
     },
     {
-      "epoch": 0.75,
-      "learning_rate": 3.759398496240601e-05,
-      "loss": 0.7102,
+      "epoch": 0.19,
+      "learning_rate": 9.345794392523365e-06,
+      "loss": 1.0983,
       "step": 100
     },
     {
-      "epoch": 0.82,
-      "learning_rate": 4.135338345864662e-05,
-      "loss": 0.6395,
+      "epoch": 0.21,
+      "learning_rate": 1.02803738317757e-05,
+      "loss": 1.1016,
       "step": 110
     },
     {
-      "epoch": 0.9,
-      "learning_rate": 4.511278195488722e-05,
-      "loss": 0.6145,
+      "epoch": 0.22,
+      "learning_rate": 1.1214953271028037e-05,
+      "loss": 1.0129,
       "step": 120
     },
     {
-      "epoch": 0.97,
-      "learning_rate": 4.887218045112782e-05,
-      "loss": 0.5832,
+      "epoch": 0.24,
+      "learning_rate": 1.2149532710280374e-05,
+      "loss": 1.0331,
       "step": 130
     },
     {
-      "epoch": 0.99,
-      "eval_accuracy": 0.8299065420560747,
-      "eval_loss": 0.5485312342643738,
-      "eval_runtime": 73.2179,
-      "eval_samples_per_second": 58.456,
-      "eval_steps_per_second": 1.83,
-      "step": 133
-    },
-    {
-      "epoch": 1.05,
-      "learning_rate": 4.970760233918128e-05,
-      "loss": 0.5927,
+      "epoch": 0.26,
+      "learning_rate": 1.308411214953271e-05,
+      "loss": 0.9059,
       "step": 140
     },
     {
-      "epoch": 1.13,
-      "learning_rate": 4.928989139515455e-05,
-      "loss": 0.4566,
+      "epoch": 0.28,
+      "learning_rate": 1.4018691588785047e-05,
+      "loss": 0.9411,
       "step": 150
     },
     {
-      "epoch": 1.2,
-      "learning_rate": 4.887218045112782e-05,
-      "loss": 0.5143,
+      "epoch": 0.3,
+      "learning_rate": 1.4953271028037382e-05,
+      "loss": 0.8249,
       "step": 160
     },
     {
-      "epoch": 1.28,
-      "learning_rate": 4.8454469507101085e-05,
-      "loss": 0.4993,
+      "epoch": 0.32,
+      "learning_rate": 1.588785046728972e-05,
+      "loss": 0.8729,
       "step": 170
     },
     {
-      "epoch": 1.35,
-      "learning_rate": 4.803675856307435e-05,
-      "loss": 0.4912,
+      "epoch": 0.34,
+      "learning_rate": 1.6822429906542056e-05,
+      "loss": 0.8896,
       "step": 180
     },
     {
-      "epoch": 1.43,
-      "learning_rate": 4.761904761904762e-05,
-      "loss": 0.4608,
+      "epoch": 0.36,
+      "learning_rate": 1.775700934579439e-05,
+      "loss": 0.8313,
       "step": 190
     },
     {
-      "epoch": 1.5,
-      "learning_rate": 4.720133667502089e-05,
-      "loss": 0.4569,
+      "epoch": 0.37,
+      "learning_rate": 1.869158878504673e-05,
+      "loss": 0.7836,
       "step": 200
     },
     {
-      "epoch": 1.58,
-      "learning_rate": 4.678362573099415e-05,
-      "loss": 0.4649,
+      "epoch": 0.39,
+      "learning_rate": 1.9626168224299065e-05,
+      "loss": 0.7587,
       "step": 210
     },
     {
-      "epoch": 1.65,
-      "learning_rate": 4.6365914786967416e-05,
-      "loss": 0.4428,
+      "epoch": 0.41,
+      "learning_rate": 2.05607476635514e-05,
+      "loss": 0.6824,
       "step": 220
     },
     {
-      "epoch": 1.73,
-      "learning_rate": 4.5948203842940684e-05,
-      "loss": 0.4654,
+      "epoch": 0.43,
+      "learning_rate": 2.149532710280374e-05,
+      "loss": 0.7291,
       "step": 230
     },
     {
-      "epoch": 1.8,
-      "learning_rate": 4.553049289891395e-05,
-      "loss": 0.4404,
+      "epoch": 0.45,
+      "learning_rate": 2.2429906542056075e-05,
+      "loss": 0.7577,
       "step": 240
     },
     {
-      "epoch": 1.87,
-      "learning_rate": 4.511278195488722e-05,
-      "loss": 0.4231,
+      "epoch": 0.47,
+      "learning_rate": 2.3364485981308414e-05,
+      "loss": 0.7538,
       "step": 250
     },
     {
-      "epoch": 1.95,
-      "learning_rate": 4.4695071010860486e-05,
-      "loss": 0.4638,
+      "epoch": 0.49,
+      "learning_rate": 2.429906542056075e-05,
+      "loss": 0.6234,
       "step": 260
     },
     {
-      "epoch": 1.99,
-      "eval_accuracy": 0.8574766355140186,
-      "eval_loss": 0.44362181425094604,
-      "eval_runtime": 75.8626,
-      "eval_samples_per_second": 56.418,
-      "eval_steps_per_second": 1.766,
-      "step": 266
-    },
-    {
-      "epoch": 2.03,
-      "learning_rate": 4.4277360066833754e-05,
-      "loss": 0.4435,
+      "epoch": 0.5,
+      "learning_rate": 2.5233644859813084e-05,
+      "loss": 0.6444,
       "step": 270
     },
     {
-      "epoch": 2.1,
-      "learning_rate": 4.3859649122807014e-05,
-      "loss": 0.3596,
+      "epoch": 0.52,
+      "learning_rate": 2.616822429906542e-05,
+      "loss": 0.6348,
       "step": 280
     },
     {
-      "epoch": 2.18,
-      "learning_rate": 4.344193817878028e-05,
-      "loss": 0.3644,
+      "epoch": 0.54,
+      "learning_rate": 2.7102803738317755e-05,
+      "loss": 0.6086,
       "step": 290
     },
     {
-      "epoch": 2.25,
-      "learning_rate": 4.302422723475355e-05,
-      "loss": 0.3731,
+      "epoch": 0.56,
+      "learning_rate": 2.8037383177570094e-05,
+      "loss": 0.6258,
       "step": 300
     },
     {
-      "epoch": 2.33,
-      "learning_rate": 4.260651629072682e-05,
-      "loss": 0.3682,
+      "epoch": 0.58,
+      "learning_rate": 2.897196261682243e-05,
+      "loss": 0.6055,
       "step": 310
     },
     {
-      "epoch": 2.4,
-      "learning_rate": 4.2188805346700084e-05,
-      "loss": 0.3619,
+      "epoch": 0.6,
+      "learning_rate": 2.9906542056074764e-05,
+      "loss": 0.5749,
       "step": 320
     },
     {
-      "epoch": 2.48,
-      "learning_rate": 4.177109440267335e-05,
-      "loss": 0.3729,
+      "epoch": 0.62,
+      "learning_rate": 3.08411214953271e-05,
+      "loss": 0.595,
       "step": 330
     },
     {
-      "epoch": 2.55,
-      "learning_rate": 4.135338345864662e-05,
-      "loss": 0.4054,
+      "epoch": 0.64,
+      "learning_rate": 3.177570093457944e-05,
+      "loss": 0.541,
       "step": 340
     },
     {
-      "epoch": 2.63,
-      "learning_rate": 4.093567251461988e-05,
-      "loss": 0.339,
+      "epoch": 0.65,
+      "learning_rate": 3.2710280373831774e-05,
+      "loss": 0.597,
       "step": 350
     },
     {
-      "epoch": 2.7,
-      "learning_rate": 4.051796157059315e-05,
-      "loss": 0.3598,
+      "epoch": 0.67,
+      "learning_rate": 3.364485981308411e-05,
+      "loss": 0.465,
       "step": 360
     },
     {
-      "epoch": 2.78,
-      "learning_rate": 4.0100250626566415e-05,
-      "loss": 0.3417,
+      "epoch": 0.69,
+      "learning_rate": 3.457943925233645e-05,
+      "loss": 0.5728,
       "step": 370
     },
     {
-      "epoch": 2.85,
-      "learning_rate": 3.968253968253968e-05,
-      "loss": 0.3379,
+      "epoch": 0.71,
+      "learning_rate": 3.551401869158878e-05,
+      "loss": 0.5068,
       "step": 380
     },
     {
-      "epoch": 2.93,
-      "learning_rate": 3.926482873851295e-05,
-      "loss": 0.3115,
+      "epoch": 0.73,
+      "learning_rate": 3.644859813084112e-05,
+      "loss": 0.5928,
       "step": 390
     },
     {
-      "epoch": 2.99,
-      "eval_accuracy": 0.8644859813084113,
-      "eval_loss": 0.4172683358192444,
-      "eval_runtime": 70.472,
-      "eval_samples_per_second": 60.733,
-      "eval_steps_per_second": 1.901,
-      "step": 399
-    },
-    {
-      "epoch": 3.01,
-      "learning_rate": 3.884711779448622e-05,
-      "loss": 0.4131,
+      "epoch": 0.75,
+      "learning_rate": 3.738317757009346e-05,
+      "loss": 0.5488,
       "step": 400
     },
     {
-      "epoch": 3.08,
-      "learning_rate": 3.8429406850459485e-05,
-      "loss": 0.2941,
+      "epoch": 0.77,
+      "learning_rate": 3.831775700934579e-05,
+      "loss": 0.5565,
       "step": 410
     },
     {
-      "epoch": 3.16,
-      "learning_rate": 3.8011695906432746e-05,
-      "loss": 0.3021,
+      "epoch": 0.79,
+      "learning_rate": 3.925233644859813e-05,
+      "loss": 0.5147,
       "step": 420
     },
     {
-      "epoch": 3.23,
-      "learning_rate": 3.759398496240601e-05,
-      "loss": 0.2967,
+      "epoch": 0.8,
+      "learning_rate": 4.018691588785047e-05,
+      "loss": 0.5709,
       "step": 430
     },
     {
-      "epoch": 3.31,
-      "learning_rate": 3.717627401837928e-05,
-      "loss": 0.2906,
+      "epoch": 0.82,
+      "learning_rate": 4.11214953271028e-05,
+      "loss": 0.5086,
       "step": 440
     },
     {
-      "epoch": 3.38,
-      "learning_rate": 3.675856307435255e-05,
-      "loss": 0.2741,
+      "epoch": 0.84,
+      "learning_rate": 4.205607476635514e-05,
+      "loss": 0.4941,
       "step": 450
     },
     {
-      "epoch": 3.46,
-      "learning_rate": 3.6340852130325816e-05,
-      "loss": 0.3069,
+      "epoch": 0.86,
+      "learning_rate": 4.299065420560748e-05,
+      "loss": 0.492,
       "step": 460
     },
     {
-      "epoch": 3.53,
-      "learning_rate": 3.592314118629908e-05,
-      "loss": 0.2697,
+      "epoch": 0.88,
+      "learning_rate": 4.392523364485982e-05,
+      "loss": 0.522,
       "step": 470
     },
     {
-      "epoch": 3.61,
-      "learning_rate": 3.5505430242272344e-05,
-      "loss": 0.2777,
+      "epoch": 0.9,
+      "learning_rate": 4.485981308411215e-05,
+      "loss": 0.5502,
       "step": 480
     },
     {
-      "epoch": 3.68,
-      "learning_rate": 3.508771929824561e-05,
-      "loss": 0.2863,
+      "epoch": 0.92,
+      "learning_rate": 4.579439252336449e-05,
+      "loss": 0.5002,
       "step": 490
     },
     {
-      "epoch": 3.76,
-      "learning_rate": 3.467000835421888e-05,
-      "loss": 0.3172,
+      "epoch": 0.93,
+      "learning_rate": 4.672897196261683e-05,
+      "loss": 0.5522,
       "step": 500
     },
     {
-      "epoch": 3.83,
-      "learning_rate": 3.4252297410192146e-05,
-      "loss": 0.2982,
+      "epoch": 0.95,
+      "learning_rate": 4.766355140186916e-05,
+      "loss": 0.5166,
       "step": 510
     },
     {
-      "epoch": 3.9,
-      "learning_rate": 3.3834586466165414e-05,
-      "loss": 0.2902,
+      "epoch": 0.97,
+      "learning_rate": 4.85981308411215e-05,
+      "loss": 0.5825,
       "step": 520
     },
     {
-      "epoch": 3.98,
-      "learning_rate": 3.341687552213868e-05,
-      "loss": 0.2926,
+      "epoch": 0.99,
+      "learning_rate": 4.9532710280373836e-05,
+      "loss": 0.5531,
       "step": 530
     },
     {
-      "epoch": 3.99,
-      "eval_accuracy": 0.8476635514018691,
-      "eval_loss": 0.44750651717185974,
-      "eval_runtime": 69.0564,
-      "eval_samples_per_second": 61.978,
-      "eval_steps_per_second": 1.94,
-      "step": 532
+      "epoch": 1.0,
+      "eval_accuracy": 0.8336448598130841,
+      "eval_loss": 0.4937918782234192,
+      "eval_runtime": 96.0141,
+      "eval_samples_per_second": 44.577,
+      "eval_steps_per_second": 5.572,
+      "step": 535
     },
     {
-      "epoch": 4.06,
-      "learning_rate": 3.299916457811195e-05,
-      "loss": 0.2541,
+      "epoch": 1.01,
+      "learning_rate": 4.994807892004154e-05,
+      "loss": 0.491,
       "step": 540
     },
     {
-      "epoch": 4.13,
-      "learning_rate": 3.258145363408521e-05,
-      "loss": 0.2287,
+      "epoch": 1.03,
+      "learning_rate": 4.9844236760124614e-05,
+      "loss": 0.4679,
       "step": 550
     },
     {
-      "epoch": 4.21,
-      "learning_rate": 3.216374269005848e-05,
-      "loss": 0.248,
+      "epoch": 1.05,
+      "learning_rate": 4.974039460020769e-05,
+      "loss": 0.4584,
       "step": 560
     },
     {
-      "epoch": 4.28,
-      "learning_rate": 3.1746031746031745e-05,
-      "loss": 0.2502,
+      "epoch": 1.07,
+      "learning_rate": 4.963655244029076e-05,
+      "loss": 0.5556,
       "step": 570
     },
     {
-      "epoch": 4.36,
-      "learning_rate": 3.132832080200501e-05,
-      "loss": 0.2053,
+      "epoch": 1.08,
+      "learning_rate": 4.9532710280373836e-05,
+      "loss": 0.4374,
       "step": 580
     },
     {
-      "epoch": 4.43,
-      "learning_rate": 3.091060985797828e-05,
-      "loss": 0.2359,
+      "epoch": 1.1,
+      "learning_rate": 4.9428868120456904e-05,
+      "loss": 0.4499,
       "step": 590
     },
     {
-      "epoch": 4.51,
-      "learning_rate": 3.0492898913951544e-05,
-      "loss": 0.2538,
+      "epoch": 1.12,
+      "learning_rate": 4.9325025960539985e-05,
+      "loss": 0.5201,
       "step": 600
     },
     {
-      "epoch": 4.58,
-      "learning_rate": 3.007518796992481e-05,
-      "loss": 0.2371,
+      "epoch": 1.14,
+      "learning_rate": 4.922118380062305e-05,
+      "loss": 0.6573,
       "step": 610
     },
     {
-      "epoch": 4.66,
-      "learning_rate": 2.965747702589808e-05,
-      "loss": 0.2402,
+      "epoch": 1.16,
+      "learning_rate": 4.9117341640706127e-05,
+      "loss": 0.5029,
       "step": 620
     },
     {
-      "epoch": 4.73,
-      "learning_rate": 2.9239766081871346e-05,
-      "loss": 0.2418,
+      "epoch": 1.18,
+      "learning_rate": 4.901349948078921e-05,
+      "loss": 0.5457,
       "step": 630
     },
     {
-      "epoch": 4.81,
-      "learning_rate": 2.882205513784461e-05,
-      "loss": 0.2465,
+      "epoch": 1.2,
+      "learning_rate": 4.8909657320872275e-05,
+      "loss": 0.492,
       "step": 640
     },
     {
-      "epoch": 4.88,
-      "learning_rate": 2.8404344193817878e-05,
-      "loss": 0.2241,
+      "epoch": 1.21,
+      "learning_rate": 4.880581516095535e-05,
+      "loss": 0.3407,
       "step": 650
     },
     {
-      "epoch": 4.96,
-      "learning_rate": 2.7986633249791145e-05,
-      "loss": 0.2127,
+      "epoch": 1.23,
+      "learning_rate": 4.8701973001038423e-05,
+      "loss": 0.5129,
       "step": 660
     },
     {
-      "epoch": 4.99,
-      "eval_accuracy": 0.8574766355140186,
-      "eval_loss": 0.4497027099132538,
-      "eval_runtime": 74.3222,
-      "eval_samples_per_second": 57.587,
-      "eval_steps_per_second": 1.803,
-      "step": 665
-    },
-    {
-      "epoch": 5.04,
-      "learning_rate": 2.756892230576441e-05,
-      "loss": 0.2127,
+      "epoch": 1.25,
+      "learning_rate": 4.85981308411215e-05,
+      "loss": 0.4626,
       "step": 670
     },
     {
-      "epoch": 5.11,
-      "learning_rate": 2.7151211361737677e-05,
-      "loss": 0.1995,
+      "epoch": 1.27,
+      "learning_rate": 4.849428868120457e-05,
+      "loss": 0.5528,
       "step": 680
     },
     {
-      "epoch": 5.19,
-      "learning_rate": 2.6733500417710944e-05,
-      "loss": 0.184,
+      "epoch": 1.29,
+      "learning_rate": 4.8390446521287646e-05,
+      "loss": 0.4977,
       "step": 690
     },
     {
-      "epoch": 5.26,
-      "learning_rate": 2.6315789473684212e-05,
-      "loss": 0.1686,
+      "epoch": 1.31,
+      "learning_rate": 4.828660436137072e-05,
+      "loss": 0.387,
       "step": 700
     },
     {
-      "epoch": 5.34,
-      "learning_rate": 2.5898078529657476e-05,
-      "loss": 0.1944,
+      "epoch": 1.33,
+      "learning_rate": 4.818276220145379e-05,
+      "loss": 0.5349,
       "step": 710
     },
     {
-      "epoch": 5.41,
-      "learning_rate": 2.5480367585630744e-05,
-      "loss": 0.1827,
+      "epoch": 1.35,
+      "learning_rate": 4.807892004153687e-05,
+      "loss": 0.4265,
       "step": 720
     },
     {
-      "epoch": 5.49,
-      "learning_rate": 2.506265664160401e-05,
-      "loss": 0.2056,
+      "epoch": 1.36,
+      "learning_rate": 4.797507788161994e-05,
+      "loss": 0.4719,
       "step": 730
     },
     {
-      "epoch": 5.56,
-      "learning_rate": 2.4644945697577275e-05,
-      "loss": 0.1706,
+      "epoch": 1.38,
+      "learning_rate": 4.787123572170301e-05,
+      "loss": 0.5005,
       "step": 740
     },
     {
-      "epoch": 5.64,
-      "learning_rate": 2.4227234753550543e-05,
-      "loss": 0.1585,
+      "epoch": 1.4,
+      "learning_rate": 4.776739356178609e-05,
+      "loss": 0.5112,
       "step": 750
     },
     {
-      "epoch": 5.71,
-      "learning_rate": 2.380952380952381e-05,
-      "loss": 0.1805,
+      "epoch": 1.42,
+      "learning_rate": 4.766355140186916e-05,
+      "loss": 0.5112,
       "step": 760
     },
     {
-      "epoch": 5.79,
-      "learning_rate": 2.3391812865497074e-05,
-      "loss": 0.2051,
+      "epoch": 1.44,
+      "learning_rate": 4.755970924195223e-05,
+      "loss": 0.3827,
       "step": 770
     },
     {
-      "epoch": 5.86,
-      "learning_rate": 2.2974101921470342e-05,
-      "loss": 0.1997,
+      "epoch": 1.46,
+      "learning_rate": 4.745586708203531e-05,
+      "loss": 0.5576,
       "step": 780
     },
     {
-      "epoch": 5.93,
-      "learning_rate": 2.255639097744361e-05,
-      "loss": 0.1934,
+      "epoch": 1.48,
+      "learning_rate": 4.735202492211838e-05,
+      "loss": 0.3974,
       "step": 790
     },
     {
-      "epoch": 5.99,
-      "eval_accuracy": 0.858177570093458,
-      "eval_loss": 0.4547787010669708,
-      "eval_runtime": 69.4508,
-      "eval_samples_per_second": 61.626,
-      "eval_steps_per_second": 1.929,
-      "step": 798
-    },
-    {
-      "epoch": 6.01,
-      "learning_rate": 2.2138680033416877e-05,
-      "loss": 0.2103,
+      "epoch": 1.5,
+      "learning_rate": 4.7248182762201456e-05,
+      "loss": 0.5209,
       "step": 800
     },
     {
-      "epoch": 6.09,
-      "learning_rate": 2.172096908939014e-05,
-      "loss": 0.1442,
+      "epoch": 1.51,
+      "learning_rate": 4.714434060228453e-05,
+      "loss": 0.4721,
       "step": 810
     },
     {
-      "epoch": 6.16,
-      "learning_rate": 2.130325814536341e-05,
-      "loss": 0.144,
+      "epoch": 1.53,
+      "learning_rate": 4.7040498442367604e-05,
+      "loss": 0.385,
       "step": 820
     },
     {
-      "epoch": 6.24,
-      "learning_rate": 2.0885547201336676e-05,
-      "loss": 0.1844,
+      "epoch": 1.55,
+      "learning_rate": 4.693665628245067e-05,
+      "loss": 0.4874,
       "step": 830
     },
     {
-      "epoch": 6.31,
-      "learning_rate": 2.046783625730994e-05,
-      "loss": 0.1574,
+      "epoch": 1.57,
+      "learning_rate": 4.683281412253375e-05,
+      "loss": 0.4409,
       "step": 840
     },
     {
-      "epoch": 6.39,
-      "learning_rate": 2.0050125313283208e-05,
-      "loss": 0.1594,
+      "epoch": 1.59,
+      "learning_rate": 4.672897196261683e-05,
+      "loss": 0.506,
       "step": 850
     },
     {
-      "epoch": 6.46,
-      "learning_rate": 1.9632414369256475e-05,
-      "loss": 0.1394,
+      "epoch": 1.61,
+      "learning_rate": 4.6625129802699895e-05,
+      "loss": 0.4446,
       "step": 860
     },
     {
-      "epoch": 6.54,
-      "learning_rate": 1.9214703425229743e-05,
-      "loss": 0.1556,
+      "epoch": 1.63,
+      "learning_rate": 4.6521287642782976e-05,
+      "loss": 0.4086,
       "step": 870
     },
     {
-      "epoch": 6.61,
-      "learning_rate": 1.8796992481203007e-05,
-      "loss": 0.1181,
+      "epoch": 1.64,
+      "learning_rate": 4.641744548286604e-05,
+      "loss": 0.3869,
       "step": 880
     },
     {
-      "epoch": 6.69,
-      "learning_rate": 1.8379281537176274e-05,
-      "loss": 0.1504,
+      "epoch": 1.66,
+      "learning_rate": 4.631360332294912e-05,
+      "loss": 0.395,
       "step": 890
     },
     {
-      "epoch": 6.76,
-      "learning_rate": 1.796157059314954e-05,
-      "loss": 0.1296,
+      "epoch": 1.68,
+      "learning_rate": 4.620976116303219e-05,
+      "loss": 0.464,
       "step": 900
     },
     {
-      "epoch": 6.84,
-      "learning_rate": 1.7543859649122806e-05,
-      "loss": 0.1444,
+      "epoch": 1.7,
+      "learning_rate": 4.6105919003115266e-05,
+      "loss": 0.3581,
       "step": 910
     },
     {
-      "epoch": 6.91,
-      "learning_rate": 1.7126148705096073e-05,
-      "loss": 0.1789,
+      "epoch": 1.72,
+      "learning_rate": 4.600207684319834e-05,
+      "loss": 0.387,
       "step": 920
     },
     {
-      "epoch": 6.99,
-      "learning_rate": 1.670843776106934e-05,
-      "loss": 0.1777,
+      "epoch": 1.74,
+      "learning_rate": 4.5898234683281414e-05,
+      "loss": 0.4066,
       "step": 930
     },
     {
-      "epoch": 6.99,
-      "eval_accuracy": 0.8560747663551402,
-      "eval_loss": 0.46795374155044556,
-      "eval_runtime": 69.8404,
-      "eval_samples_per_second": 61.283,
-      "eval_steps_per_second": 1.919,
-      "step": 931
-    },
-    {
-      "epoch": 7.07,
-      "learning_rate": 1.6290726817042605e-05,
-      "loss": 0.132,
+      "epoch": 1.76,
+      "learning_rate": 4.579439252336449e-05,
+      "loss": 0.3997,
       "step": 940
     },
     {
-      "epoch": 7.14,
-      "learning_rate": 1.5873015873015872e-05,
-      "loss": 0.1483,
+      "epoch": 1.78,
+      "learning_rate": 4.569055036344756e-05,
+      "loss": 0.4705,
       "step": 950
     },
     {
-      "epoch": 7.22,
-      "learning_rate": 1.545530492898914e-05,
-      "loss": 0.1322,
+      "epoch": 1.79,
+      "learning_rate": 4.558670820353064e-05,
+      "loss": 0.4063,
       "step": 960
     },
     {
-      "epoch": 7.29,
-      "learning_rate": 1.5037593984962406e-05,
-      "loss": 0.1382,
+      "epoch": 1.81,
+      "learning_rate": 4.548286604361371e-05,
+      "loss": 0.4227,
       "step": 970
     },
     {
-      "epoch": 7.37,
-      "learning_rate": 1.4619883040935673e-05,
-      "loss": 0.1041,
+      "epoch": 1.83,
+      "learning_rate": 4.537902388369678e-05,
+      "loss": 0.4844,
       "step": 980
     },
     {
-      "epoch": 7.44,
-      "learning_rate": 1.4202172096908939e-05,
-      "loss": 0.096,
+      "epoch": 1.85,
+      "learning_rate": 4.527518172377986e-05,
+      "loss": 0.4631,
       "step": 990
     },
     {
-      "epoch": 7.52,
-      "learning_rate": 1.3784461152882205e-05,
-      "loss": 0.1336,
+      "epoch": 1.87,
+      "learning_rate": 4.517133956386293e-05,
+      "loss": 0.4847,
       "step": 1000
     },
     {
-      "epoch": 7.59,
-      "learning_rate": 1.3366750208855472e-05,
-      "loss": 0.1269,
+      "epoch": 1.89,
+      "learning_rate": 4.5067497403946e-05,
+      "loss": 0.4112,
       "step": 1010
     },
     {
-      "epoch": 7.67,
-      "learning_rate": 1.2949039264828738e-05,
-      "loss": 0.116,
+      "epoch": 1.91,
+      "learning_rate": 4.496365524402908e-05,
+      "loss": 0.3964,
       "step": 1020
     },
     {
-      "epoch": 7.74,
-      "learning_rate": 1.2531328320802006e-05,
-      "loss": 0.1386,
+      "epoch": 1.93,
+      "learning_rate": 4.485981308411215e-05,
+      "loss": 0.4341,
       "step": 1030
     },
     {
-      "epoch": 7.81,
-      "learning_rate": 1.2113617376775271e-05,
-      "loss": 0.1228,
+      "epoch": 1.94,
+      "learning_rate": 4.4755970924195224e-05,
+      "loss": 0.5106,
       "step": 1040
     },
     {
-      "epoch": 7.89,
-      "learning_rate": 1.1695906432748537e-05,
-      "loss": 0.103,
+      "epoch": 1.96,
+      "learning_rate": 4.46521287642783e-05,
+      "loss": 0.4269,
       "step": 1050
     },
     {
-      "epoch": 7.96,
-      "learning_rate": 1.1278195488721805e-05,
-      "loss": 0.1187,
+      "epoch": 1.98,
+      "learning_rate": 4.454828660436137e-05,
+      "loss": 0.413,
       "step": 1060
     },
     {
-      "epoch": 7.99,
-      "eval_accuracy": 0.8591121495327103,
-      "eval_loss": 0.4879584014415741,
-      "eval_runtime": 69.5594,
-      "eval_samples_per_second": 61.53,
-      "eval_steps_per_second": 1.926,
-      "step": 1064
+      "epoch": 2.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4139,
+      "step": 1070
     },
     {
-      "epoch": 8.04,
-      "learning_rate": 1.086048454469507e-05,
-      "loss": 0.1422,
+      "epoch": 2.0,
+      "eval_accuracy": 0.861214953271028,
+      "eval_loss": 0.40707165002822876,
+      "eval_runtime": 99.8719,
+      "eval_samples_per_second": 42.855,
+      "eval_steps_per_second": 5.357,
       "step": 1070
     },
     {
-      "epoch": 8.12,
-      "learning_rate": 1.0442773600668338e-05,
-      "loss": 0.0984,
+      "epoch": 2.02,
+      "learning_rate": 4.434060228452752e-05,
+      "loss": 0.395,
       "step": 1080
     },
     {
-      "epoch": 8.19,
-      "learning_rate": 1.0025062656641604e-05,
-      "loss": 0.1016,
+      "epoch": 2.04,
+      "learning_rate": 4.4236760124610595e-05,
+      "loss": 0.3672,
       "step": 1090
     },
     {
-      "epoch": 8.27,
-      "learning_rate": 9.607351712614871e-06,
-      "loss": 0.097,
+      "epoch": 2.06,
+      "learning_rate": 4.413291796469366e-05,
+      "loss": 0.3994,
       "step": 1100
     },
     {
-      "epoch": 8.34,
-      "learning_rate": 9.189640768588137e-06,
-      "loss": 0.1138,
+      "epoch": 2.07,
+      "learning_rate": 4.4029075804776743e-05,
+      "loss": 0.3818,
       "step": 1110
     },
     {
-      "epoch": 8.42,
-      "learning_rate": 8.771929824561403e-06,
-      "loss": 0.0806,
+      "epoch": 2.09,
+      "learning_rate": 4.392523364485982e-05,
+      "loss": 0.4013,
       "step": 1120
     },
     {
-      "epoch": 8.49,
-      "learning_rate": 8.35421888053467e-06,
-      "loss": 0.0968,
+      "epoch": 2.11,
+      "learning_rate": 4.3821391484942885e-05,
+      "loss": 0.3409,
       "step": 1130
     },
     {
-      "epoch": 8.57,
-      "learning_rate": 7.936507936507936e-06,
-      "loss": 0.0855,
+      "epoch": 2.13,
+      "learning_rate": 4.3717549325025966e-05,
+      "loss": 0.4753,
       "step": 1140
     },
     {
-      "epoch": 8.64,
-      "learning_rate": 7.518796992481203e-06,
-      "loss": 0.1295,
+      "epoch": 2.15,
+      "learning_rate": 4.3613707165109034e-05,
+      "loss": 0.3671,
       "step": 1150
     },
     {
-      "epoch": 8.72,
-      "learning_rate": 7.1010860484544695e-06,
-      "loss": 0.1062,
+      "epoch": 2.17,
+      "learning_rate": 4.350986500519211e-05,
+      "loss": 0.462,
       "step": 1160
     },
     {
-      "epoch": 8.79,
-      "learning_rate": 6.683375104427736e-06,
-      "loss": 0.0931,
+      "epoch": 2.19,
+      "learning_rate": 4.340602284527518e-05,
+      "loss": 0.3729,
       "step": 1170
     },
     {
-      "epoch": 8.87,
-      "learning_rate": 6.265664160401003e-06,
-      "loss": 0.118,
+      "epoch": 2.21,
+      "learning_rate": 4.3302180685358256e-05,
+      "loss": 0.3876,
       "step": 1180
     },
     {
-      "epoch": 8.94,
-      "learning_rate": 5.8479532163742686e-06,
-      "loss": 0.0801,
+      "epoch": 2.22,
+      "learning_rate": 4.319833852544133e-05,
+      "loss": 0.3794,
       "step": 1190
     },
     {
-      "epoch": 8.99,
-      "eval_accuracy": 0.8556074766355141,
-      "eval_loss": 0.5013762712478638,
-      "eval_runtime": 69.6837,
-      "eval_samples_per_second": 61.42,
-      "eval_steps_per_second": 1.923,
-      "step": 1197
-    },
-    {
-      "epoch": 9.02,
-      "learning_rate": 5.430242272347535e-06,
-      "loss": 0.1025,
+      "epoch": 2.24,
+      "learning_rate": 4.3094496365524405e-05,
+      "loss": 0.3609,
       "step": 1200
     },
     {
-      "epoch": 9.1,
-      "learning_rate": 5.012531328320802e-06,
-      "loss": 0.0854,
+      "epoch": 2.26,
+      "learning_rate": 4.299065420560748e-05,
+      "loss": 0.4161,
       "step": 1210
     },
     {
-      "epoch": 9.17,
-      "learning_rate": 4.5948203842940685e-06,
-      "loss": 0.1032,
+      "epoch": 2.28,
+      "learning_rate": 4.2886812045690546e-05,
+      "loss": 0.5202,
       "step": 1220
     },
     {
-      "epoch": 9.25,
-      "learning_rate": 4.177109440267335e-06,
-      "loss": 0.108,
+      "epoch": 2.3,
+      "learning_rate": 4.278296988577363e-05,
+      "loss": 0.4579,
       "step": 1230
     },
     {
-      "epoch": 9.32,
-      "learning_rate": 3.7593984962406014e-06,
-      "loss": 0.0933,
+      "epoch": 2.32,
+      "learning_rate": 4.26791277258567e-05,
+      "loss": 0.408,
       "step": 1240
     },
     {
-      "epoch": 9.4,
-      "learning_rate": 3.341687552213868e-06,
-      "loss": 0.0795,
+      "epoch": 2.34,
+      "learning_rate": 4.257528556593977e-05,
+      "loss": 0.36,
       "step": 1250
     },
     {
-      "epoch": 9.47,
-      "learning_rate": 2.9239766081871343e-06,
-      "loss": 0.0882,
+      "epoch": 2.36,
+      "learning_rate": 4.247144340602285e-05,
+      "loss": 0.3109,
       "step": 1260
     },
     {
-      "epoch": 9.55,
-      "learning_rate": 2.506265664160401e-06,
-      "loss": 0.0907,
+      "epoch": 2.37,
+      "learning_rate": 4.236760124610592e-05,
+      "loss": 0.3713,
       "step": 1270
     },
     {
-      "epoch": 9.62,
-      "learning_rate": 2.0885547201336676e-06,
-      "loss": 0.0775,
+      "epoch": 2.39,
+      "learning_rate": 4.226375908618899e-05,
+      "loss": 0.4329,
       "step": 1280
     },
     {
-      "epoch": 9.7,
-      "learning_rate": 1.670843776106934e-06,
-      "loss": 0.091,
+      "epoch": 2.41,
+      "learning_rate": 4.2159916926272066e-05,
+      "loss": 0.3612,
       "step": 1290
     },
     {
-      "epoch": 9.77,
-      "learning_rate": 1.2531328320802005e-06,
-      "loss": 0.0741,
+      "epoch": 2.43,
+      "learning_rate": 4.205607476635514e-05,
+      "loss": 0.387,
       "step": 1300
     },
     {
-      "epoch": 9.84,
-      "learning_rate": 8.35421888053467e-07,
-      "loss": 0.0728,
+      "epoch": 2.45,
+      "learning_rate": 4.1952232606438215e-05,
+      "loss": 0.4461,
       "step": 1310
     },
     {
-      "epoch": 9.92,
-      "learning_rate": 4.177109440267335e-07,
-      "loss": 0.1036,
+      "epoch": 2.47,
+      "learning_rate": 4.184839044652129e-05,
+      "loss": 0.2903,
       "step": 1320
     },
     {
-      "epoch": 9.99,
-      "learning_rate": 0.0,
-      "loss": 0.088,
+      "epoch": 2.49,
+      "learning_rate": 4.174454828660436e-05,
+      "loss": 0.328,
       "step": 1330
     },
     {
-      "epoch": 9.99,
-      "eval_accuracy": 0.8558411214953271,
-      "eval_loss": 0.5052832365036011,
-      "eval_runtime": 69.3744,
-      "eval_samples_per_second": 61.694,
-      "eval_steps_per_second": 1.932,
-      "step": 1330
+      "epoch": 2.5,
+      "learning_rate": 4.164070612668744e-05,
+      "loss": 0.4151,
+      "step": 1340
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 4.153686396677051e-05,
+      "loss": 0.3888,
+      "step": 1350
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 4.1433021806853586e-05,
+      "loss": 0.3871,
+      "step": 1360
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 4.132917964693666e-05,
+      "loss": 0.4304,
+      "step": 1370
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 4.1225337487019734e-05,
+      "loss": 0.387,
+      "step": 1380
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.11214953271028e-05,
+      "loss": 0.3658,
+      "step": 1390
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.101765316718588e-05,
+      "loss": 0.4494,
+      "step": 1400
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 4.091381100726896e-05,
+      "loss": 0.3355,
+      "step": 1410
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 4.0809968847352024e-05,
+      "loss": 0.3368,
+      "step": 1420
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.0706126687435105e-05,
+      "loss": 0.4087,
+      "step": 1430
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 4.060228452751817e-05,
+      "loss": 0.4279,
+      "step": 1440
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.049844236760125e-05,
+      "loss": 0.3548,
+      "step": 1450
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 4.039460020768432e-05,
+      "loss": 0.4466,
+      "step": 1460
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.0290758047767395e-05,
+      "loss": 0.4133,
+      "step": 1470
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.018691588785047e-05,
+      "loss": 0.3845,
+      "step": 1480
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 4.0083073727933544e-05,
+      "loss": 0.4471,
+      "step": 1490
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.997923156801662e-05,
+      "loss": 0.4421,
+      "step": 1500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3.987538940809969e-05,
+      "loss": 0.4383,
+      "step": 1510
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.9771547248182767e-05,
+      "loss": 0.3326,
+      "step": 1520
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 3.966770508826584e-05,
+      "loss": 0.4058,
+      "step": 1530
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 3.956386292834891e-05,
+      "loss": 0.3653,
+      "step": 1540
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.946002076843199e-05,
+      "loss": 0.4157,
+      "step": 1550
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 3.935617860851506e-05,
+      "loss": 0.3944,
+      "step": 1560
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.925233644859813e-05,
+      "loss": 0.4103,
+      "step": 1570
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.914849428868121e-05,
+      "loss": 0.4601,
+      "step": 1580
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.904465212876428e-05,
+      "loss": 0.3561,
+      "step": 1590
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.8940809968847354e-05,
+      "loss": 0.287,
+      "step": 1600
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8642523364485981,
+      "eval_loss": 0.39544418454170227,
+      "eval_runtime": 95.8501,
+      "eval_samples_per_second": 44.653,
+      "eval_steps_per_second": 5.582,
+      "step": 1605
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.883696780893043e-05,
+      "loss": 0.3781,
+      "step": 1610
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.87331256490135e-05,
+      "loss": 0.3999,
+      "step": 1620
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.8629283489096576e-05,
+      "loss": 0.3507,
+      "step": 1630
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.852544132917965e-05,
+      "loss": 0.354,
+      "step": 1640
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.8421599169262725e-05,
+      "loss": 0.4233,
+      "step": 1650
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.831775700934579e-05,
+      "loss": 0.3766,
+      "step": 1660
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.821391484942887e-05,
+      "loss": 0.4494,
+      "step": 1670
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.811007268951195e-05,
+      "loss": 0.3772,
+      "step": 1680
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.8006230529595015e-05,
+      "loss": 0.4079,
+      "step": 1690
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.7902388369678096e-05,
+      "loss": 0.3796,
+      "step": 1700
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.779854620976116e-05,
+      "loss": 0.2899,
+      "step": 1710
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.769470404984424e-05,
+      "loss": 0.357,
+      "step": 1720
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.759086188992731e-05,
+      "loss": 0.3922,
+      "step": 1730
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.7487019730010386e-05,
+      "loss": 0.3743,
+      "step": 1740
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.738317757009346e-05,
+      "loss": 0.3275,
+      "step": 1750
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.7279335410176535e-05,
+      "loss": 0.3128,
+      "step": 1760
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.717549325025961e-05,
+      "loss": 0.3351,
+      "step": 1770
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.7071651090342676e-05,
+      "loss": 0.354,
+      "step": 1780
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.696780893042576e-05,
+      "loss": 0.3604,
+      "step": 1790
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.686396677050883e-05,
+      "loss": 0.389,
+      "step": 1800
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.67601246105919e-05,
+      "loss": 0.3892,
+      "step": 1810
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.665628245067498e-05,
+      "loss": 0.3525,
+      "step": 1820
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.655244029075805e-05,
+      "loss": 0.3725,
+      "step": 1830
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.644859813084112e-05,
+      "loss": 0.3909,
+      "step": 1840
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.6344755970924196e-05,
+      "loss": 0.2884,
+      "step": 1850
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.624091381100727e-05,
+      "loss": 0.3384,
+      "step": 1860
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.6137071651090344e-05,
+      "loss": 0.3645,
+      "step": 1870
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.603322949117342e-05,
+      "loss": 0.3407,
+      "step": 1880
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.592938733125649e-05,
+      "loss": 0.3847,
+      "step": 1890
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.582554517133957e-05,
+      "loss": 0.3862,
+      "step": 1900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.572170301142264e-05,
+      "loss": 0.3831,
+      "step": 1910
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.5617860851505715e-05,
+      "loss": 0.3147,
+      "step": 1920
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.551401869158878e-05,
+      "loss": 0.3686,
+      "step": 1930
     },
     {
-      "epoch": 9.99,
-      "step": 1330,
-      "total_flos": 1.3257453564799912e+19,
-      "train_loss": 0.3000902961967583,
-      "train_runtime": 6004.4248,
-      "train_samples_per_second": 28.507,
-      "train_steps_per_second": 0.222
+      "epoch": 3.63,
+      "learning_rate": 3.5410176531671864e-05,
+      "loss": 0.3913,
+      "step": 1940
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.530633437175493e-05,
+      "loss": 0.3717,
+      "step": 1950
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.5202492211838006e-05,
+      "loss": 0.4135,
+      "step": 1960
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.5098650051921087e-05,
+      "loss": 0.372,
+      "step": 1970
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.4994807892004154e-05,
+      "loss": 0.3738,
+      "step": 1980
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.489096573208723e-05,
+      "loss": 0.3847,
+      "step": 1990
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.47871235721703e-05,
+      "loss": 0.3893,
+      "step": 2000
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.468328141225338e-05,
+      "loss": 0.3045,
+      "step": 2010
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.457943925233645e-05,
+      "loss": 0.3242,
+      "step": 2020
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.4475597092419525e-05,
+      "loss": 0.3417,
+      "step": 2030
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.43717549325026e-05,
+      "loss": 0.3568,
+      "step": 2040
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.426791277258567e-05,
+      "loss": 0.3225,
+      "step": 2050
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.416407061266875e-05,
+      "loss": 0.3341,
+      "step": 2060
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.406022845275182e-05,
+      "loss": 0.326,
+      "step": 2070
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.395638629283489e-05,
+      "loss": 0.3508,
+      "step": 2080
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.385254413291797e-05,
+      "loss": 0.333,
+      "step": 2090
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.374870197300104e-05,
+      "loss": 0.4373,
+      "step": 2100
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.364485981308411e-05,
+      "loss": 0.3769,
+      "step": 2110
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.3541017653167186e-05,
+      "loss": 0.3697,
+      "step": 2120
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.343717549325026e-05,
+      "loss": 0.3187,
+      "step": 2130
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4211,
+      "step": 2140
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8700934579439252,
+      "eval_loss": 0.39060959219932556,
+      "eval_runtime": 93.1195,
+      "eval_samples_per_second": 45.962,
+      "eval_steps_per_second": 5.745,
+      "step": 2140
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 3.322949117341641e-05,
+      "loss": 0.2857,
+      "step": 2150
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 3.3125649013499483e-05,
+      "loss": 0.3271,
+      "step": 2160
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 3.302180685358255e-05,
+      "loss": 0.3109,
+      "step": 2170
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.291796469366563e-05,
+      "loss": 0.2679,
+      "step": 2180
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 3.2814122533748706e-05,
+      "loss": 0.3242,
+      "step": 2190
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 3.2710280373831774e-05,
+      "loss": 0.3083,
+      "step": 2200
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 3.2606438213914855e-05,
+      "loss": 0.3325,
+      "step": 2210
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 3.250259605399792e-05,
+      "loss": 0.3989,
+      "step": 2220
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 3.2398753894080996e-05,
+      "loss": 0.3044,
+      "step": 2230
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 3.229491173416408e-05,
+      "loss": 0.3389,
+      "step": 2240
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 3.2191069574247145e-05,
+      "loss": 0.3284,
+      "step": 2250
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 3.208722741433022e-05,
+      "loss": 0.2777,
+      "step": 2260
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 3.198338525441329e-05,
+      "loss": 0.3531,
+      "step": 2270
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 3.187954309449637e-05,
+      "loss": 0.3578,
+      "step": 2280
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 3.177570093457944e-05,
+      "loss": 0.375,
+      "step": 2290
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 3.1671858774662516e-05,
+      "loss": 0.3912,
+      "step": 2300
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 3.156801661474559e-05,
+      "loss": 0.3632,
+      "step": 2310
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 3.146417445482866e-05,
+      "loss": 0.3453,
+      "step": 2320
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 3.136033229491174e-05,
+      "loss": 0.3412,
+      "step": 2330
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 3.1256490134994806e-05,
+      "loss": 0.3506,
+      "step": 2340
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 3.115264797507788e-05,
+      "loss": 0.3178,
+      "step": 2350
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 3.104880581516096e-05,
+      "loss": 0.3787,
+      "step": 2360
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.094496365524403e-05,
+      "loss": 0.2522,
+      "step": 2370
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 3.08411214953271e-05,
+      "loss": 0.2986,
+      "step": 2380
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 3.073727933541018e-05,
+      "loss": 0.3647,
+      "step": 2390
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 3.063343717549325e-05,
+      "loss": 0.2927,
+      "step": 2400
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 3.0529595015576326e-05,
+      "loss": 0.3868,
+      "step": 2410
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 3.04257528556594e-05,
+      "loss": 0.2927,
+      "step": 2420
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 3.0321910695742474e-05,
+      "loss": 0.2755,
+      "step": 2430
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 3.0218068535825545e-05,
+      "loss": 0.2953,
+      "step": 2440
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.0114226375908622e-05,
+      "loss": 0.2946,
+      "step": 2450
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.0010384215991693e-05,
+      "loss": 0.3302,
+      "step": 2460
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 2.9906542056074764e-05,
+      "loss": 0.406,
+      "step": 2470
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.9802699896157842e-05,
+      "loss": 0.323,
+      "step": 2480
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.9698857736240916e-05,
+      "loss": 0.3808,
+      "step": 2490
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.9595015576323987e-05,
+      "loss": 0.3269,
+      "step": 2500
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.9491173416407064e-05,
+      "loss": 0.3703,
+      "step": 2510
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 2.9387331256490135e-05,
+      "loss": 0.3173,
+      "step": 2520
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 2.9283489096573206e-05,
+      "loss": 0.2749,
+      "step": 2530
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.9179646936656284e-05,
+      "loss": 0.3128,
+      "step": 2540
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 2.9075804776739358e-05,
+      "loss": 0.3725,
+      "step": 2550
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 2.897196261682243e-05,
+      "loss": 0.3052,
+      "step": 2560
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 2.8868120456905506e-05,
+      "loss": 0.3159,
+      "step": 2570
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.8764278296988577e-05,
+      "loss": 0.27,
+      "step": 2580
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.866043613707165e-05,
+      "loss": 0.3566,
+      "step": 2590
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.855659397715473e-05,
+      "loss": 0.3622,
+      "step": 2600
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.84527518172378e-05,
+      "loss": 0.3318,
+      "step": 2610
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.834890965732087e-05,
+      "loss": 0.334,
+      "step": 2620
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.824506749740395e-05,
+      "loss": 0.3084,
+      "step": 2630
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.814122533748702e-05,
+      "loss": 0.3524,
+      "step": 2640
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.8037383177570094e-05,
+      "loss": 0.3473,
+      "step": 2650
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.793354101765317e-05,
+      "loss": 0.3593,
+      "step": 2660
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.7829698857736242e-05,
+      "loss": 0.316,
+      "step": 2670
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.8754672897196262,
+      "eval_loss": 0.37160804867744446,
+      "eval_runtime": 95.1891,
+      "eval_samples_per_second": 44.963,
+      "eval_steps_per_second": 5.62,
+      "step": 2675
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 2.7725856697819313e-05,
+      "loss": 0.2738,
+      "step": 2680
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.762201453790239e-05,
+      "loss": 0.342,
+      "step": 2690
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 2.751817237798546e-05,
+      "loss": 0.3448,
+      "step": 2700
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.7414330218068536e-05,
+      "loss": 0.3099,
+      "step": 2710
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 2.7310488058151613e-05,
+      "loss": 0.3368,
+      "step": 2720
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.7206645898234684e-05,
+      "loss": 0.2743,
+      "step": 2730
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 2.7102803738317755e-05,
+      "loss": 0.3064,
+      "step": 2740
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 2.6998961578400832e-05,
+      "loss": 0.2109,
+      "step": 2750
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.6895119418483907e-05,
+      "loss": 0.3611,
+      "step": 2760
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 2.6791277258566978e-05,
+      "loss": 0.3456,
+      "step": 2770
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.6687435098650055e-05,
+      "loss": 0.2599,
+      "step": 2780
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.6583592938733126e-05,
+      "loss": 0.2951,
+      "step": 2790
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.6479750778816197e-05,
+      "loss": 0.1794,
+      "step": 2800
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2.6375908618899274e-05,
+      "loss": 0.2954,
+      "step": 2810
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.627206645898235e-05,
+      "loss": 0.298,
+      "step": 2820
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 2.616822429906542e-05,
+      "loss": 0.3187,
+      "step": 2830
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 2.6064382139148497e-05,
+      "loss": 0.3297,
+      "step": 2840
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.5960539979231568e-05,
+      "loss": 0.2431,
+      "step": 2850
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 2.585669781931464e-05,
+      "loss": 0.3967,
+      "step": 2860
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 2.5752855659397716e-05,
+      "loss": 0.3722,
+      "step": 2870
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 2.564901349948079e-05,
+      "loss": 0.2752,
+      "step": 2880
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.554517133956386e-05,
+      "loss": 0.2971,
+      "step": 2890
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 2.544132917964694e-05,
+      "loss": 0.2804,
+      "step": 2900
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.533748701973001e-05,
+      "loss": 0.3715,
+      "step": 2910
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 2.5233644859813084e-05,
+      "loss": 0.2833,
+      "step": 2920
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 2.512980269989616e-05,
+      "loss": 0.2891,
+      "step": 2930
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.5025960539979233e-05,
+      "loss": 0.3122,
+      "step": 2940
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 2.4922118380062307e-05,
+      "loss": 0.2891,
+      "step": 2950
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.481827622014538e-05,
+      "loss": 0.3167,
+      "step": 2960
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 2.4714434060228452e-05,
+      "loss": 0.3505,
+      "step": 2970
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.4610591900311526e-05,
+      "loss": 0.3255,
+      "step": 2980
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 2.4506749740394604e-05,
+      "loss": 0.2708,
+      "step": 2990
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 2.4402907580477675e-05,
+      "loss": 0.3709,
+      "step": 3000
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.429906542056075e-05,
+      "loss": 0.3623,
+      "step": 3010
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 2.4195223260643823e-05,
+      "loss": 0.2488,
+      "step": 3020
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 2.4091381100726894e-05,
+      "loss": 0.3012,
+      "step": 3030
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 2.398753894080997e-05,
+      "loss": 0.318,
+      "step": 3040
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.3883696780893046e-05,
+      "loss": 0.3596,
+      "step": 3050
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 2.3779854620976117e-05,
+      "loss": 0.2923,
+      "step": 3060
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.367601246105919e-05,
+      "loss": 0.3677,
+      "step": 3070
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 2.3572170301142265e-05,
+      "loss": 0.2508,
+      "step": 3080
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 2.3468328141225336e-05,
+      "loss": 0.2464,
+      "step": 3090
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.3364485981308414e-05,
+      "loss": 0.3364,
+      "step": 3100
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.3260643821391488e-05,
+      "loss": 0.2931,
+      "step": 3110
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.315680166147456e-05,
+      "loss": 0.2962,
+      "step": 3120
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.3052959501557633e-05,
+      "loss": 0.282,
+      "step": 3130
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.2949117341640707e-05,
+      "loss": 0.2978,
+      "step": 3140
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 2.284527518172378e-05,
+      "loss": 0.2524,
+      "step": 3150
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 2.2741433021806856e-05,
+      "loss": 0.3384,
+      "step": 3160
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.263759086188993e-05,
+      "loss": 0.2936,
+      "step": 3170
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 2.2533748701973e-05,
+      "loss": 0.2521,
+      "step": 3180
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.2429906542056075e-05,
+      "loss": 0.3685,
+      "step": 3190
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 2.232606438213915e-05,
+      "loss": 0.2391,
+      "step": 3200
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.2709,
+      "step": 3210
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.8735981308411215,
+      "eval_loss": 0.37839841842651367,
+      "eval_runtime": 92.127,
+      "eval_samples_per_second": 46.458,
+      "eval_steps_per_second": 5.807,
+      "step": 3210
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 2.2118380062305298e-05,
+      "loss": 0.2305,
+      "step": 3220
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 2.2014537902388372e-05,
+      "loss": 0.2758,
+      "step": 3230
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 2.1910695742471443e-05,
+      "loss": 0.2602,
+      "step": 3240
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 2.1806853582554517e-05,
+      "loss": 0.311,
+      "step": 3250
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 2.170301142263759e-05,
+      "loss": 0.3315,
+      "step": 3260
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 2.1599169262720665e-05,
+      "loss": 0.3078,
+      "step": 3270
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.149532710280374e-05,
+      "loss": 0.319,
+      "step": 3280
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 2.1391484942886814e-05,
+      "loss": 0.3326,
+      "step": 3290
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 2.1287642782969885e-05,
+      "loss": 0.2892,
+      "step": 3300
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 2.118380062305296e-05,
+      "loss": 0.2669,
+      "step": 3310
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 2.1079958463136033e-05,
+      "loss": 0.2829,
+      "step": 3320
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 2.0976116303219107e-05,
+      "loss": 0.2434,
+      "step": 3330
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 2.087227414330218e-05,
+      "loss": 0.2422,
+      "step": 3340
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 2.0768431983385256e-05,
+      "loss": 0.3655,
+      "step": 3350
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 2.066458982346833e-05,
+      "loss": 0.2772,
+      "step": 3360
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 2.05607476635514e-05,
+      "loss": 0.2542,
+      "step": 3370
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 2.045690550363448e-05,
+      "loss": 0.2978,
+      "step": 3380
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 2.0353063343717553e-05,
+      "loss": 0.2518,
+      "step": 3390
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 2.0249221183800623e-05,
+      "loss": 0.2697,
+      "step": 3400
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 2.0145379023883698e-05,
+      "loss": 0.2326,
+      "step": 3410
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 2.0041536863966772e-05,
+      "loss": 0.29,
+      "step": 3420
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 1.9937694704049846e-05,
+      "loss": 0.2858,
+      "step": 3430
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 1.983385254413292e-05,
+      "loss": 0.2665,
+      "step": 3440
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 1.9730010384215995e-05,
+      "loss": 0.2583,
+      "step": 3450
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 1.9626168224299065e-05,
+      "loss": 0.3213,
+      "step": 3460
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 1.952232606438214e-05,
+      "loss": 0.3385,
+      "step": 3470
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 1.9418483904465214e-05,
+      "loss": 0.2492,
+      "step": 3480
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 1.9314641744548288e-05,
+      "loss": 0.2829,
+      "step": 3490
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 1.9210799584631362e-05,
+      "loss": 0.2651,
+      "step": 3500
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.9106957424714437e-05,
+      "loss": 0.2671,
+      "step": 3510
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 1.9003115264797507e-05,
+      "loss": 0.3116,
+      "step": 3520
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 1.889927310488058e-05,
+      "loss": 0.3116,
+      "step": 3530
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 1.8795430944963656e-05,
+      "loss": 0.2305,
+      "step": 3540
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 1.869158878504673e-05,
+      "loss": 0.3035,
+      "step": 3550
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.8587746625129804e-05,
+      "loss": 0.2388,
+      "step": 3560
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 1.848390446521288e-05,
+      "loss": 0.2424,
+      "step": 3570
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.838006230529595e-05,
+      "loss": 0.219,
+      "step": 3580
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 1.8276220145379024e-05,
+      "loss": 0.212,
+      "step": 3590
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 1.8172377985462098e-05,
+      "loss": 0.3682,
+      "step": 3600
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 1.8068535825545172e-05,
+      "loss": 0.2882,
+      "step": 3610
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 1.7964693665628246e-05,
+      "loss": 0.2846,
+      "step": 3620
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 1.786085150571132e-05,
+      "loss": 0.276,
+      "step": 3630
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 1.775700934579439e-05,
+      "loss": 0.214,
+      "step": 3640
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 1.7653167185877466e-05,
+      "loss": 0.2351,
+      "step": 3650
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 1.7549325025960543e-05,
+      "loss": 0.2428,
+      "step": 3660
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 1.7445482866043614e-05,
+      "loss": 0.2289,
+      "step": 3670
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.734164070612669e-05,
+      "loss": 0.2087,
+      "step": 3680
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 1.7237798546209763e-05,
+      "loss": 0.2492,
+      "step": 3690
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 1.7133956386292833e-05,
+      "loss": 0.2855,
+      "step": 3700
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.703011422637591e-05,
+      "loss": 0.2701,
+      "step": 3710
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 1.6926272066458985e-05,
+      "loss": 0.2635,
+      "step": 3720
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 1.6822429906542056e-05,
+      "loss": 0.267,
+      "step": 3730
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.671858774662513e-05,
+      "loss": 0.177,
+      "step": 3740
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.8745327102803738,
+      "eval_loss": 0.37715595960617065,
+      "eval_runtime": 95.1073,
+      "eval_samples_per_second": 45.002,
+      "eval_steps_per_second": 5.625,
+      "step": 3745
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 1.6614745586708205e-05,
+      "loss": 0.307,
+      "step": 3750
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 1.6510903426791275e-05,
+      "loss": 0.258,
+      "step": 3760
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.6407061266874353e-05,
+      "loss": 0.2322,
+      "step": 3770
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 1.6303219106957427e-05,
+      "loss": 0.187,
+      "step": 3780
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 1.6199376947040498e-05,
+      "loss": 0.2838,
+      "step": 3790
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 1.6095534787123572e-05,
+      "loss": 0.2353,
+      "step": 3800
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 1.5991692627206647e-05,
+      "loss": 0.2661,
+      "step": 3810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 1.588785046728972e-05,
+      "loss": 0.1917,
+      "step": 3820
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 1.5784008307372795e-05,
+      "loss": 0.1978,
+      "step": 3830
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 1.568016614745587e-05,
+      "loss": 0.2688,
+      "step": 3840
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.557632398753894e-05,
+      "loss": 0.2383,
+      "step": 3850
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.5472481827622014e-05,
+      "loss": 0.2874,
+      "step": 3860
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.536863966770509e-05,
+      "loss": 0.265,
+      "step": 3870
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.5264797507788163e-05,
+      "loss": 0.1821,
+      "step": 3880
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.5160955347871237e-05,
+      "loss": 0.193,
+      "step": 3890
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.5057113187954311e-05,
+      "loss": 0.2681,
+      "step": 3900
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.4953271028037382e-05,
+      "loss": 0.2406,
+      "step": 3910
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.4849428868120458e-05,
+      "loss": 0.3214,
+      "step": 3920
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.4745586708203532e-05,
+      "loss": 0.2521,
+      "step": 3930
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.4641744548286603e-05,
+      "loss": 0.3427,
+      "step": 3940
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.4537902388369679e-05,
+      "loss": 0.3272,
+      "step": 3950
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.4434060228452753e-05,
+      "loss": 0.2336,
+      "step": 3960
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.4330218068535826e-05,
+      "loss": 0.2209,
+      "step": 3970
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.42263759086189e-05,
+      "loss": 0.222,
+      "step": 3980
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.4122533748701974e-05,
+      "loss": 0.236,
+      "step": 3990
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.4018691588785047e-05,
+      "loss": 0.2875,
+      "step": 4000
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.3914849428868121e-05,
+      "loss": 0.2098,
+      "step": 4010
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.3811007268951195e-05,
+      "loss": 0.324,
+      "step": 4020
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.3707165109034268e-05,
+      "loss": 0.2038,
+      "step": 4030
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.3603322949117342e-05,
+      "loss": 0.271,
+      "step": 4040
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.3499480789200416e-05,
+      "loss": 0.296,
+      "step": 4050
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.3395638629283489e-05,
+      "loss": 0.2557,
+      "step": 4060
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.3291796469366563e-05,
+      "loss": 0.2788,
+      "step": 4070
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.3187954309449637e-05,
+      "loss": 0.2684,
+      "step": 4080
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.308411214953271e-05,
+      "loss": 0.2447,
+      "step": 4090
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.2980269989615784e-05,
+      "loss": 0.287,
+      "step": 4100
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.2876427829698858e-05,
+      "loss": 0.2186,
+      "step": 4110
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.277258566978193e-05,
+      "loss": 0.1978,
+      "step": 4120
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.2668743509865005e-05,
+      "loss": 0.2698,
+      "step": 4130
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.256490134994808e-05,
+      "loss": 0.1624,
+      "step": 4140
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.2461059190031153e-05,
+      "loss": 0.2591,
+      "step": 4150
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.2357217030114226e-05,
+      "loss": 0.2035,
+      "step": 4160
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.2253374870197302e-05,
+      "loss": 0.2285,
+      "step": 4170
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.2149532710280374e-05,
+      "loss": 0.2257,
+      "step": 4180
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.2045690550363447e-05,
+      "loss": 0.2555,
+      "step": 4190
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.1941848390446523e-05,
+      "loss": 0.2192,
+      "step": 4200
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.1838006230529595e-05,
+      "loss": 0.2169,
+      "step": 4210
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.1734164070612668e-05,
+      "loss": 0.2914,
+      "step": 4220
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.1630321910695744e-05,
+      "loss": 0.253,
+      "step": 4230
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.1526479750778816e-05,
+      "loss": 0.2067,
+      "step": 4240
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.142263759086189e-05,
+      "loss": 0.1995,
+      "step": 4250
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.1318795430944965e-05,
+      "loss": 0.2902,
+      "step": 4260
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.1214953271028037e-05,
+      "loss": 0.2051,
+      "step": 4270
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 0.2409,
+      "step": 4280
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8761682242990654,
+      "eval_loss": 0.38753223419189453,
+      "eval_runtime": 91.8924,
+      "eval_samples_per_second": 46.576,
+      "eval_steps_per_second": 5.822,
+      "step": 4280
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.1007268951194186e-05,
+      "loss": 0.2436,
+      "step": 4290
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.0903426791277258e-05,
+      "loss": 0.2468,
+      "step": 4300
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.0799584631360333e-05,
+      "loss": 0.168,
+      "step": 4310
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.0695742471443407e-05,
+      "loss": 0.2093,
+      "step": 4320
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.059190031152648e-05,
+      "loss": 0.1616,
+      "step": 4330
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.0488058151609554e-05,
+      "loss": 0.1912,
+      "step": 4340
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.0384215991692628e-05,
+      "loss": 0.3157,
+      "step": 4350
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.02803738317757e-05,
+      "loss": 0.2213,
+      "step": 4360
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.0176531671858776e-05,
+      "loss": 0.2473,
+      "step": 4370
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.0072689511941849e-05,
+      "loss": 0.2155,
+      "step": 4380
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 9.968847352024923e-06,
+      "loss": 0.209,
+      "step": 4390
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 9.865005192107997e-06,
+      "loss": 0.2552,
+      "step": 4400
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 9.76116303219107e-06,
+      "loss": 0.2469,
+      "step": 4410
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 9.657320872274144e-06,
+      "loss": 0.2195,
+      "step": 4420
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 9.553478712357218e-06,
+      "loss": 0.2094,
+      "step": 4430
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 9.44963655244029e-06,
+      "loss": 0.2593,
+      "step": 4440
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 9.345794392523365e-06,
+      "loss": 0.254,
+      "step": 4450
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 9.24195223260644e-06,
+      "loss": 0.1353,
+      "step": 4460
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 9.138110072689512e-06,
+      "loss": 0.2152,
+      "step": 4470
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 9.034267912772586e-06,
+      "loss": 0.2258,
+      "step": 4480
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 8.93042575285566e-06,
+      "loss": 0.1593,
+      "step": 4490
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 8.826583592938733e-06,
+      "loss": 0.1939,
+      "step": 4500
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 8.722741433021807e-06,
+      "loss": 0.2717,
+      "step": 4510
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 8.618899273104881e-06,
+      "loss": 0.2263,
+      "step": 4520
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 8.515057113187956e-06,
+      "loss": 0.281,
+      "step": 4530
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 8.411214953271028e-06,
+      "loss": 0.2438,
+      "step": 4540
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 8.307372793354102e-06,
+      "loss": 0.1487,
+      "step": 4550
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 8.203530633437177e-06,
+      "loss": 0.1811,
+      "step": 4560
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 8.099688473520249e-06,
+      "loss": 0.2803,
+      "step": 4570
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 7.995846313603323e-06,
+      "loss": 0.2339,
+      "step": 4580
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 7.892004153686398e-06,
+      "loss": 0.2617,
+      "step": 4590
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 7.78816199376947e-06,
+      "loss": 0.3019,
+      "step": 4600
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 7.684319833852544e-06,
+      "loss": 0.2083,
+      "step": 4610
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 7.5804776739356185e-06,
+      "loss": 0.24,
+      "step": 4620
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 7.476635514018691e-06,
+      "loss": 0.2378,
+      "step": 4630
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 7.372793354101766e-06,
+      "loss": 0.2475,
+      "step": 4640
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 7.2689511941848395e-06,
+      "loss": 0.197,
+      "step": 4650
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 7.165109034267913e-06,
+      "loss": 0.2302,
+      "step": 4660
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 7.061266874350987e-06,
+      "loss": 0.2186,
+      "step": 4670
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 6.9574247144340605e-06,
+      "loss": 0.2172,
+      "step": 4680
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 6.853582554517134e-06,
+      "loss": 0.2189,
+      "step": 4690
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 6.749740394600208e-06,
+      "loss": 0.2137,
+      "step": 4700
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 6.6458982346832815e-06,
+      "loss": 0.2259,
+      "step": 4710
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 6.542056074766355e-06,
+      "loss": 0.15,
+      "step": 4720
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.438213914849429e-06,
+      "loss": 0.2622,
+      "step": 4730
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.3343717549325025e-06,
+      "loss": 0.2009,
+      "step": 4740
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.230529595015577e-06,
+      "loss": 0.2095,
+      "step": 4750
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.126687435098651e-06,
+      "loss": 0.1766,
+      "step": 4760
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.0228452751817235e-06,
+      "loss": 0.1982,
+      "step": 4770
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.919003115264798e-06,
+      "loss": 0.2287,
+      "step": 4780
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.815160955347872e-06,
+      "loss": 0.266,
+      "step": 4790
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.711318795430945e-06,
+      "loss": 0.2196,
+      "step": 4800
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.607476635514019e-06,
+      "loss": 0.1929,
+      "step": 4810
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.8707943925233644,
+      "eval_loss": 0.3915020227432251,
+      "eval_runtime": 93.367,
+      "eval_samples_per_second": 45.841,
+      "eval_steps_per_second": 5.73,
+      "step": 4815
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.503634475597093e-06,
+      "loss": 0.2137,
+      "step": 4820
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 5.399792315680166e-06,
+      "loss": 0.1994,
+      "step": 4830
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 5.29595015576324e-06,
+      "loss": 0.2243,
+      "step": 4840
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 5.192107995846314e-06,
+      "loss": 0.282,
+      "step": 4850
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 5.088265835929388e-06,
+      "loss": 0.1425,
+      "step": 4860
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.9844236760124615e-06,
+      "loss": 0.2213,
+      "step": 4870
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.880581516095535e-06,
+      "loss": 0.1939,
+      "step": 4880
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.776739356178609e-06,
+      "loss": 0.1753,
+      "step": 4890
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.6728971962616825e-06,
+      "loss": 0.1758,
+      "step": 4900
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.569055036344756e-06,
+      "loss": 0.1597,
+      "step": 4910
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 4.46521287642783e-06,
+      "loss": 0.2269,
+      "step": 4920
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 4.3613707165109035e-06,
+      "loss": 0.2723,
+      "step": 4930
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 4.257528556593978e-06,
+      "loss": 0.2012,
+      "step": 4940
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 4.153686396677051e-06,
+      "loss": 0.1863,
+      "step": 4950
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 4.0498442367601245e-06,
+      "loss": 0.2029,
+      "step": 4960
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.946002076843199e-06,
+      "loss": 0.2067,
+      "step": 4970
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 3.842159916926272e-06,
+      "loss": 0.2359,
+      "step": 4980
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.7383177570093455e-06,
+      "loss": 0.2398,
+      "step": 4990
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 3.6344755970924198e-06,
+      "loss": 0.2233,
+      "step": 5000
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 3.5306334371754936e-06,
+      "loss": 0.1934,
+      "step": 5010
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 3.426791277258567e-06,
+      "loss": 0.2625,
+      "step": 5020
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 3.3229491173416407e-06,
+      "loss": 0.2548,
+      "step": 5030
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 3.2191069574247146e-06,
+      "loss": 0.1718,
+      "step": 5040
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 3.1152647975077884e-06,
+      "loss": 0.1668,
+      "step": 5050
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 3.0114226375908617e-06,
+      "loss": 0.2202,
+      "step": 5060
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 2.907580477673936e-06,
+      "loss": 0.2091,
+      "step": 5070
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 2.8037383177570094e-06,
+      "loss": 0.2213,
+      "step": 5080
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 2.699896157840083e-06,
+      "loss": 0.2857,
+      "step": 5090
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 2.596053997923157e-06,
+      "loss": 0.2209,
+      "step": 5100
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 2.4922118380062308e-06,
+      "loss": 0.2341,
+      "step": 5110
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 2.3883696780893046e-06,
+      "loss": 0.2111,
+      "step": 5120
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 2.284527518172378e-06,
+      "loss": 0.1869,
+      "step": 5130
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 2.1806853582554518e-06,
+      "loss": 0.2229,
+      "step": 5140
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 2.0768431983385256e-06,
+      "loss": 0.2251,
+      "step": 5150
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 1.9730010384215994e-06,
+      "loss": 0.183,
+      "step": 5160
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 1.8691588785046728e-06,
+      "loss": 0.1606,
+      "step": 5170
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.7653167185877468e-06,
+      "loss": 0.2067,
+      "step": 5180
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 1.6614745586708204e-06,
+      "loss": 0.2346,
+      "step": 5190
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 1.5576323987538942e-06,
+      "loss": 0.2161,
+      "step": 5200
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 1.453790238836968e-06,
+      "loss": 0.2893,
+      "step": 5210
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 1.3499480789200416e-06,
+      "loss": 0.2196,
+      "step": 5220
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 1.2461059190031154e-06,
+      "loss": 0.2035,
+      "step": 5230
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 1.142263759086189e-06,
+      "loss": 0.1541,
+      "step": 5240
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.0384215991692628e-06,
+      "loss": 0.2105,
+      "step": 5250
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 9.345794392523364e-07,
+      "loss": 0.2065,
+      "step": 5260
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 8.307372793354102e-07,
+      "loss": 0.161,
+      "step": 5270
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 7.26895119418484e-07,
+      "loss": 0.2285,
+      "step": 5280
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 6.230529595015577e-07,
+      "loss": 0.2051,
+      "step": 5290
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 5.192107995846314e-07,
+      "loss": 0.2291,
+      "step": 5300
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 4.153686396677051e-07,
+      "loss": 0.2071,
+      "step": 5310
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 3.1152647975077885e-07,
+      "loss": 0.2055,
+      "step": 5320
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 2.0768431983385255e-07,
+      "loss": 0.1433,
+      "step": 5330
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 1.0384215991692627e-07,
+      "loss": 0.2493,
+      "step": 5340
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 0.0,
+      "loss": 0.1877,
+      "step": 5350
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.8714953271028038,
+      "eval_loss": 0.38807550072669983,
+      "eval_runtime": 92.2965,
+      "eval_samples_per_second": 46.372,
+      "eval_steps_per_second": 5.797,
+      "step": 5350
+    },
+    {
+      "epoch": 10.0,
+      "step": 5350,
+      "total_flos": 1.3264660513609667e+19,
+      "train_loss": 0.36025063641717503,
+      "train_runtime": 7147.2488,
+      "train_samples_per_second": 23.949,
+      "train_steps_per_second": 0.749
     }
   ],
-  "max_steps": 1330,
+  "max_steps": 5350,
   "num_train_epochs": 10,
-  "total_flos": 1.3257453564799912e+19,
+  "total_flos": 1.3264660513609667e+19,
   "trial_name": null,
   "trial_params": null
 }