diff --git a/model.safetensors b/model.safetensors
index f3e31a4092479d56e8168f94b6b90fb17ca6e407..02b0ebd42a539c8320f98b1cf77921cac1509918 100644
--- a/model.safetensors
+++ b/model.safetensors
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fd0bdc5483d786e2b3f7c85777a9c669be94de7087b07a6141776eb245d18eb9
+oid sha256:45a8e9cf842600f235cd6591e45120a144ba7d680bf9ae4d9f92472afaa7a475
 size 343233204
diff --git a/tmp-checkpoint-3060/config.json b/tmp-checkpoint-3060/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-3060/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-3060/model.safetensors b/tmp-checkpoint-3060/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..b30fe4955af3047444147816ddfe0e03e3ceb000
--- /dev/null
+++ b/tmp-checkpoint-3060/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b40f9106dece576a12de9807992aeea4b3cac8ab45f79bb8b996e00274844a13
+size 343233204
diff --git a/tmp-checkpoint-3060/optimizer.pt b/tmp-checkpoint-3060/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..402d95bdb4c8a6cb2db1a69c4fc483e1e3ac64f6
--- /dev/null
+++ b/tmp-checkpoint-3060/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54d14902f2fa125adf1d72ac1aab157dc35e3f63a8ea86baf43844ab9a4027c7
+size 686587322
diff --git a/tmp-checkpoint-3060/preprocessor_config.json b/tmp-checkpoint-3060/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-3060/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-3060/rng_state.pth b/tmp-checkpoint-3060/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..befeaae1448806d690ca2011d2d1fa332b16551a
--- /dev/null
+++ b/tmp-checkpoint-3060/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0db111771306e648ad34526cac88d4d465008a1fe7effb063f85c5ae13bda237
+size 14244
diff --git a/tmp-checkpoint-3060/scheduler.pt b/tmp-checkpoint-3060/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ee7773aef50e8012290b6dd3799e2842e1e9519b
--- /dev/null
+++ b/tmp-checkpoint-3060/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6362b4096d477b5540efd94c69181162ebbd80e5e3b8297c43eb6937f6870848
+size 1064
diff --git a/tmp-checkpoint-3060/trainer_state.json b/tmp-checkpoint-3060/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4442679817a6466138826178140cf043af6d558d
--- /dev/null
+++ b/tmp-checkpoint-3060/trainer_state.json
@@ -0,0 +1,2097 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 12.0,
+  "eval_steps": 500,
+  "global_step": 3060,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 1.8970549125031526e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-3060/training_args.bin b/tmp-checkpoint-3060/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-3060/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-3315/config.json b/tmp-checkpoint-3315/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-3315/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-3315/model.safetensors b/tmp-checkpoint-3315/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..b6f8210900f78e3db3ce37c09a74cde1184fcc45
--- /dev/null
+++ b/tmp-checkpoint-3315/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8e8465902ef580e0fdd59815d1e48ed0ad9832de44245ad73cdfd4fa08e4955
+size 343233204
diff --git a/tmp-checkpoint-3315/optimizer.pt b/tmp-checkpoint-3315/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..efda3a2780aaf1b8bd5486f4ae1af46d4c349ee8
--- /dev/null
+++ b/tmp-checkpoint-3315/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47c5f57d854d71c921ba96258797f38f48ca735d95c851544ae3dda0b1db35d0
+size 686587322
diff --git a/tmp-checkpoint-3315/preprocessor_config.json b/tmp-checkpoint-3315/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-3315/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-3315/rng_state.pth b/tmp-checkpoint-3315/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d41fd8e86da4f203afd49dcf59930737c4134cdb
--- /dev/null
+++ b/tmp-checkpoint-3315/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5bf7e96bddc2abbab56b0e243dc2d471df256e5f3ad1d04217705d81461a84f
+size 14244
diff --git a/tmp-checkpoint-3315/scheduler.pt b/tmp-checkpoint-3315/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ce2419097d064475e82ded688f46accb3321eb05
--- /dev/null
+++ b/tmp-checkpoint-3315/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15185b1efd6adbb1e200d344c648438e58b2a4f8ec52ddd433fd45fe7951a901
+size 1064
diff --git a/tmp-checkpoint-3315/trainer_state.json b/tmp-checkpoint-3315/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ea069ab789c7e3875999660ee1dbdedfdfe98034
--- /dev/null
+++ b/tmp-checkpoint-3315/trainer_state.json
@@ -0,0 +1,2267 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 13.0,
+  "eval_steps": 500,
+  "global_step": 3315,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.0551428218784154e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-3315/training_args.bin b/tmp-checkpoint-3315/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-3315/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-3570/config.json b/tmp-checkpoint-3570/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-3570/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-3570/model.safetensors b/tmp-checkpoint-3570/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..9f2b1213a72413adb95fbb77316bc7799c574d23
--- /dev/null
+++ b/tmp-checkpoint-3570/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56002304c224a74229b4d405da0f9a81ddb491427b068a69cafd3a7730a13088
+size 343233204
diff --git a/tmp-checkpoint-3570/optimizer.pt b/tmp-checkpoint-3570/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4787fbbb1d6689455950258682609622826849d6
--- /dev/null
+++ b/tmp-checkpoint-3570/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ac8dacf91028b0b511eaa766d3643d915c66dbcf54edd8bbdf59510d3b9446e
+size 686587322
diff --git a/tmp-checkpoint-3570/preprocessor_config.json b/tmp-checkpoint-3570/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-3570/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-3570/rng_state.pth b/tmp-checkpoint-3570/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..524364b8b28a998cf32254a9ad19ef5314785c18
--- /dev/null
+++ b/tmp-checkpoint-3570/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ba6791e7daac8ddc7a01505409eac2ae48ce0e7b7dcc18e60fab41ad9cc1939
+size 14244
diff --git a/tmp-checkpoint-3570/scheduler.pt b/tmp-checkpoint-3570/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e41890c12755e72e5fe6c50ade1cb41adf5efc57
--- /dev/null
+++ b/tmp-checkpoint-3570/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccbea5df28e0d7b245f97c0f296c72a00d1848b50f8903e5b8b9620596a42158
+size 1064
diff --git a/tmp-checkpoint-3570/trainer_state.json b/tmp-checkpoint-3570/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..252a1114dc70f497c7b4c7e08809342c143d8075
--- /dev/null
+++ b/tmp-checkpoint-3570/trainer_state.json
@@ -0,0 +1,2443 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 14.0,
+  "eval_steps": 500,
+  "global_step": 3570,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.1445170660856935e-05,
+      "loss": 0.1158,
+      "step": 3320
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.1086,
+      "step": 3330
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.129992737835875e-05,
+      "loss": 0.1612,
+      "step": 3340
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.122730573710966e-05,
+      "loss": 0.1272,
+      "step": 3350
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.115468409586057e-05,
+      "loss": 0.2715,
+      "step": 3360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.1082062454611474e-05,
+      "loss": 0.1432,
+      "step": 3370
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.100944081336238e-05,
+      "loss": 0.0573,
+      "step": 3380
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.093681917211329e-05,
+      "loss": 0.141,
+      "step": 3390
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.08641975308642e-05,
+      "loss": 0.156,
+      "step": 3400
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.0791575889615106e-05,
+      "loss": 0.1558,
+      "step": 3410
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.1606,
+      "step": 3420
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.064633260711692e-05,
+      "loss": 0.1677,
+      "step": 3430
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.057371096586783e-05,
+      "loss": 0.0562,
+      "step": 3440
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.0501089324618738e-05,
+      "loss": 0.1684,
+      "step": 3450
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.0428467683369643e-05,
+      "loss": 0.3589,
+      "step": 3460
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.0355846042120557e-05,
+      "loss": 0.2076,
+      "step": 3470
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.0283224400871462e-05,
+      "loss": 0.0773,
+      "step": 3480
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.021060275962237e-05,
+      "loss": 0.0424,
+      "step": 3490
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.0137981118373278e-05,
+      "loss": 0.1573,
+      "step": 3500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.0465,
+      "step": 3510
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.999273783587509e-05,
+      "loss": 0.0091,
+      "step": 3520
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.9920116194626e-05,
+      "loss": 0.2262,
+      "step": 3530
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.984749455337691e-05,
+      "loss": 0.0253,
+      "step": 3540
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.9774872912127817e-05,
+      "loss": 0.184,
+      "step": 3550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.9702251270878722e-05,
+      "loss": 0.1289,
+      "step": 3560
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.2017,
+      "step": 3570
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9274509803921569
+      },
+      "eval_f1": {
+        "f1": 0.9275152403725876
+      },
+      "eval_loss": 0.3324253261089325,
+      "eval_precision": {
+        "precision": 0.9309834626869261
+      },
+      "eval_recall": {
+        "recall": 0.9275085729872963
+      },
+      "eval_runtime": 10.2109,
+      "eval_samples_per_second": 49.947,
+      "eval_steps_per_second": 24.973,
+      "step": 3570
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.213230731253678e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-3570/training_args.bin b/tmp-checkpoint-3570/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-3570/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-3825/config.json b/tmp-checkpoint-3825/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-3825/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-3825/model.safetensors b/tmp-checkpoint-3825/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..477df644c227d6d590acb6bf078ee3c12cdfd3c1
--- /dev/null
+++ b/tmp-checkpoint-3825/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b908fbbb00a78b9995aff82f3624d66b569be281979e5e625be16b9f3af5fd93
+size 343233204
diff --git a/tmp-checkpoint-3825/optimizer.pt b/tmp-checkpoint-3825/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..69ddb05ee37508d72b6c4118d32d2c092586f081
--- /dev/null
+++ b/tmp-checkpoint-3825/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8de4140a42b3bec09f4d4a6be569463c354ba1d735a35847d03886546114926
+size 686587322
diff --git a/tmp-checkpoint-3825/preprocessor_config.json b/tmp-checkpoint-3825/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-3825/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-3825/rng_state.pth b/tmp-checkpoint-3825/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..6661df8dc8fbb7640ac5cbcdc6d62fb00ef429d8
--- /dev/null
+++ b/tmp-checkpoint-3825/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b57a281fb50a3ddf84c43ba0739dc4e9a4a0688bf7bfa1cbf259ba8756714b3
+size 14244
diff --git a/tmp-checkpoint-3825/scheduler.pt b/tmp-checkpoint-3825/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..cbeb17343366ff387c181045fbacc8269ecbe0fb
--- /dev/null
+++ b/tmp-checkpoint-3825/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b99e487a626d6199525e9fd65b4aba946d5708b2e350fc1552f0dfdebf9f5270
+size 1064
diff --git a/tmp-checkpoint-3825/trainer_state.json b/tmp-checkpoint-3825/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..bf97930681b7f67c9b12027c57633406e1b2d7c2
--- /dev/null
+++ b/tmp-checkpoint-3825/trainer_state.json
@@ -0,0 +1,2613 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 15.0,
+  "eval_steps": 500,
+  "global_step": 3825,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.1445170660856935e-05,
+      "loss": 0.1158,
+      "step": 3320
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.1086,
+      "step": 3330
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.129992737835875e-05,
+      "loss": 0.1612,
+      "step": 3340
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.122730573710966e-05,
+      "loss": 0.1272,
+      "step": 3350
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.115468409586057e-05,
+      "loss": 0.2715,
+      "step": 3360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.1082062454611474e-05,
+      "loss": 0.1432,
+      "step": 3370
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.100944081336238e-05,
+      "loss": 0.0573,
+      "step": 3380
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.093681917211329e-05,
+      "loss": 0.141,
+      "step": 3390
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.08641975308642e-05,
+      "loss": 0.156,
+      "step": 3400
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.0791575889615106e-05,
+      "loss": 0.1558,
+      "step": 3410
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.1606,
+      "step": 3420
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.064633260711692e-05,
+      "loss": 0.1677,
+      "step": 3430
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.057371096586783e-05,
+      "loss": 0.0562,
+      "step": 3440
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.0501089324618738e-05,
+      "loss": 0.1684,
+      "step": 3450
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.0428467683369643e-05,
+      "loss": 0.3589,
+      "step": 3460
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.0355846042120557e-05,
+      "loss": 0.2076,
+      "step": 3470
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.0283224400871462e-05,
+      "loss": 0.0773,
+      "step": 3480
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.021060275962237e-05,
+      "loss": 0.0424,
+      "step": 3490
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.0137981118373278e-05,
+      "loss": 0.1573,
+      "step": 3500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.0465,
+      "step": 3510
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.999273783587509e-05,
+      "loss": 0.0091,
+      "step": 3520
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.9920116194626e-05,
+      "loss": 0.2262,
+      "step": 3530
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.984749455337691e-05,
+      "loss": 0.0253,
+      "step": 3540
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.9774872912127817e-05,
+      "loss": 0.184,
+      "step": 3550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.9702251270878722e-05,
+      "loss": 0.1289,
+      "step": 3560
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.2017,
+      "step": 3570
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9274509803921569
+      },
+      "eval_f1": {
+        "f1": 0.9275152403725876
+      },
+      "eval_loss": 0.3324253261089325,
+      "eval_precision": {
+        "precision": 0.9309834626869261
+      },
+      "eval_recall": {
+        "recall": 0.9275085729872963
+      },
+      "eval_runtime": 10.2109,
+      "eval_samples_per_second": 49.947,
+      "eval_steps_per_second": 24.973,
+      "step": 3570
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 2.9557007988380538e-05,
+      "loss": 0.2156,
+      "step": 3580
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 2.948438634713145e-05,
+      "loss": 0.1672,
+      "step": 3590
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.0842,
+      "step": 3600
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 2.9339143064633262e-05,
+      "loss": 0.1529,
+      "step": 3610
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 2.926652142338417e-05,
+      "loss": 0.0904,
+      "step": 3620
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 2.9193899782135074e-05,
+      "loss": 0.1129,
+      "step": 3630
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.912127814088599e-05,
+      "loss": 0.0721,
+      "step": 3640
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 2.9048656499636893e-05,
+      "loss": 0.137,
+      "step": 3650
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 2.89760348583878e-05,
+      "loss": 0.1637,
+      "step": 3660
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.890341321713871e-05,
+      "loss": 0.1298,
+      "step": 3670
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 2.8830791575889614e-05,
+      "loss": 0.037,
+      "step": 3680
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.1013,
+      "step": 3690
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 2.8685548293391433e-05,
+      "loss": 0.1149,
+      "step": 3700
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 2.861292665214234e-05,
+      "loss": 0.1502,
+      "step": 3710
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 2.854030501089325e-05,
+      "loss": 0.1915,
+      "step": 3720
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 2.8467683369644154e-05,
+      "loss": 0.0854,
+      "step": 3730
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 2.839506172839506e-05,
+      "loss": 0.0395,
+      "step": 3740
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.832244008714597e-05,
+      "loss": 0.1681,
+      "step": 3750
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.824981844589688e-05,
+      "loss": 0.1236,
+      "step": 3760
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.817719680464779e-05,
+      "loss": 0.1138,
+      "step": 3770
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.1743,
+      "step": 3780
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.80319535221496e-05,
+      "loss": 0.0698,
+      "step": 3790
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.795933188090051e-05,
+      "loss": 0.1649,
+      "step": 3800
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.788671023965142e-05,
+      "loss": 0.0095,
+      "step": 3810
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.781408859840233e-05,
+      "loss": 0.017,
+      "step": 3820
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.9235294117647059
+      },
+      "eval_f1": {
+        "f1": 0.9236114478832926
+      },
+      "eval_loss": 0.319390207529068,
+      "eval_precision": {
+        "precision": 0.9261443200762086
+      },
+      "eval_recall": {
+        "recall": 0.9255596796820201
+      },
+      "eval_runtime": 10.1385,
+      "eval_samples_per_second": 50.303,
+      "eval_steps_per_second": 25.152,
+      "step": 3825
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.371318640628941e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-3825/training_args.bin b/tmp-checkpoint-3825/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-3825/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-4080/config.json b/tmp-checkpoint-4080/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-4080/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-4080/model.safetensors b/tmp-checkpoint-4080/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..4a776a7379fae5e15602f522e9148450cfbf537d
--- /dev/null
+++ b/tmp-checkpoint-4080/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd9d8134928d5f172548c38b5f8f98f646d80ebe930143d6a8ce6ed66c75e171
+size 343233204
diff --git a/tmp-checkpoint-4080/optimizer.pt b/tmp-checkpoint-4080/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9920f3d8eaa3e4725b6bf55f1c66f3b54fbd8f22
--- /dev/null
+++ b/tmp-checkpoint-4080/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:397901202f25d01902acc32eddbda572c0f7d44a94c06ed8f07facbd162d700e
+size 686587322
diff --git a/tmp-checkpoint-4080/preprocessor_config.json b/tmp-checkpoint-4080/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-4080/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-4080/rng_state.pth b/tmp-checkpoint-4080/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..e12edf791f6ebc52fd494417a15725a70b95548d
--- /dev/null
+++ b/tmp-checkpoint-4080/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c82c427c5f0ebd36ed667ff7530aed732b7dc47d8cb858d2ef9e6eb786cd2cc3
+size 14244
diff --git a/tmp-checkpoint-4080/scheduler.pt b/tmp-checkpoint-4080/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8926f421efeb58e199fc5c806cbc12c2c15688ac
--- /dev/null
+++ b/tmp-checkpoint-4080/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4674d4810713ae3a231c485f455783d3c0e4d9a9ca5609127e78a842672e0436
+size 1064
diff --git a/tmp-checkpoint-4080/trainer_state.json b/tmp-checkpoint-4080/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b7efbb78240bc0e8527e6b3211705d035ecc2902
--- /dev/null
+++ b/tmp-checkpoint-4080/trainer_state.json
@@ -0,0 +1,2789 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 16.0,
+  "eval_steps": 500,
+  "global_step": 4080,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.1445170660856935e-05,
+      "loss": 0.1158,
+      "step": 3320
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.1086,
+      "step": 3330
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.129992737835875e-05,
+      "loss": 0.1612,
+      "step": 3340
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.122730573710966e-05,
+      "loss": 0.1272,
+      "step": 3350
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.115468409586057e-05,
+      "loss": 0.2715,
+      "step": 3360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.1082062454611474e-05,
+      "loss": 0.1432,
+      "step": 3370
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.100944081336238e-05,
+      "loss": 0.0573,
+      "step": 3380
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.093681917211329e-05,
+      "loss": 0.141,
+      "step": 3390
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.08641975308642e-05,
+      "loss": 0.156,
+      "step": 3400
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.0791575889615106e-05,
+      "loss": 0.1558,
+      "step": 3410
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.1606,
+      "step": 3420
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.064633260711692e-05,
+      "loss": 0.1677,
+      "step": 3430
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.057371096586783e-05,
+      "loss": 0.0562,
+      "step": 3440
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.0501089324618738e-05,
+      "loss": 0.1684,
+      "step": 3450
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.0428467683369643e-05,
+      "loss": 0.3589,
+      "step": 3460
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.0355846042120557e-05,
+      "loss": 0.2076,
+      "step": 3470
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.0283224400871462e-05,
+      "loss": 0.0773,
+      "step": 3480
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.021060275962237e-05,
+      "loss": 0.0424,
+      "step": 3490
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.0137981118373278e-05,
+      "loss": 0.1573,
+      "step": 3500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.0465,
+      "step": 3510
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.999273783587509e-05,
+      "loss": 0.0091,
+      "step": 3520
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.9920116194626e-05,
+      "loss": 0.2262,
+      "step": 3530
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.984749455337691e-05,
+      "loss": 0.0253,
+      "step": 3540
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.9774872912127817e-05,
+      "loss": 0.184,
+      "step": 3550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.9702251270878722e-05,
+      "loss": 0.1289,
+      "step": 3560
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.2017,
+      "step": 3570
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9274509803921569
+      },
+      "eval_f1": {
+        "f1": 0.9275152403725876
+      },
+      "eval_loss": 0.3324253261089325,
+      "eval_precision": {
+        "precision": 0.9309834626869261
+      },
+      "eval_recall": {
+        "recall": 0.9275085729872963
+      },
+      "eval_runtime": 10.2109,
+      "eval_samples_per_second": 49.947,
+      "eval_steps_per_second": 24.973,
+      "step": 3570
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 2.9557007988380538e-05,
+      "loss": 0.2156,
+      "step": 3580
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 2.948438634713145e-05,
+      "loss": 0.1672,
+      "step": 3590
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.0842,
+      "step": 3600
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 2.9339143064633262e-05,
+      "loss": 0.1529,
+      "step": 3610
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 2.926652142338417e-05,
+      "loss": 0.0904,
+      "step": 3620
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 2.9193899782135074e-05,
+      "loss": 0.1129,
+      "step": 3630
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.912127814088599e-05,
+      "loss": 0.0721,
+      "step": 3640
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 2.9048656499636893e-05,
+      "loss": 0.137,
+      "step": 3650
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 2.89760348583878e-05,
+      "loss": 0.1637,
+      "step": 3660
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.890341321713871e-05,
+      "loss": 0.1298,
+      "step": 3670
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 2.8830791575889614e-05,
+      "loss": 0.037,
+      "step": 3680
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.1013,
+      "step": 3690
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 2.8685548293391433e-05,
+      "loss": 0.1149,
+      "step": 3700
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 2.861292665214234e-05,
+      "loss": 0.1502,
+      "step": 3710
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 2.854030501089325e-05,
+      "loss": 0.1915,
+      "step": 3720
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 2.8467683369644154e-05,
+      "loss": 0.0854,
+      "step": 3730
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 2.839506172839506e-05,
+      "loss": 0.0395,
+      "step": 3740
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.832244008714597e-05,
+      "loss": 0.1681,
+      "step": 3750
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.824981844589688e-05,
+      "loss": 0.1236,
+      "step": 3760
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.817719680464779e-05,
+      "loss": 0.1138,
+      "step": 3770
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.1743,
+      "step": 3780
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.80319535221496e-05,
+      "loss": 0.0698,
+      "step": 3790
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.795933188090051e-05,
+      "loss": 0.1649,
+      "step": 3800
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.788671023965142e-05,
+      "loss": 0.0095,
+      "step": 3810
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.781408859840233e-05,
+      "loss": 0.017,
+      "step": 3820
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.9235294117647059
+      },
+      "eval_f1": {
+        "f1": 0.9236114478832926
+      },
+      "eval_loss": 0.319390207529068,
+      "eval_precision": {
+        "precision": 0.9261443200762086
+      },
+      "eval_recall": {
+        "recall": 0.9255596796820201
+      },
+      "eval_runtime": 10.1385,
+      "eval_samples_per_second": 50.303,
+      "eval_steps_per_second": 25.152,
+      "step": 3825
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 2.7741466957153233e-05,
+      "loss": 0.007,
+      "step": 3830
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.766884531590414e-05,
+      "loss": 0.1578,
+      "step": 3840
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 2.759622367465505e-05,
+      "loss": 0.0678,
+      "step": 3850
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 2.7523602033405954e-05,
+      "loss": 0.0545,
+      "step": 3860
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.1241,
+      "step": 3870
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.7378358750907773e-05,
+      "loss": 0.1044,
+      "step": 3880
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 2.730573710965868e-05,
+      "loss": 0.2882,
+      "step": 3890
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.7233115468409585e-05,
+      "loss": 0.1343,
+      "step": 3900
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.7160493827160493e-05,
+      "loss": 0.0315,
+      "step": 3910
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 2.70878721859114e-05,
+      "loss": 0.2315,
+      "step": 3920
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 2.7015250544662313e-05,
+      "loss": 0.0577,
+      "step": 3930
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 2.694262890341322e-05,
+      "loss": 0.1342,
+      "step": 3940
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.6870007262164125e-05,
+      "loss": 0.0996,
+      "step": 3950
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.1453,
+      "step": 3960
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 2.672476397966594e-05,
+      "loss": 0.0898,
+      "step": 3970
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.6652142338416852e-05,
+      "loss": 0.0078,
+      "step": 3980
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 2.657952069716776e-05,
+      "loss": 0.0463,
+      "step": 3990
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 2.6506899055918665e-05,
+      "loss": 0.1273,
+      "step": 4000
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.6434277414669573e-05,
+      "loss": 0.0209,
+      "step": 4010
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.636165577342048e-05,
+      "loss": 0.0307,
+      "step": 4020
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.6289034132171385e-05,
+      "loss": 0.1268,
+      "step": 4030
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.62164124909223e-05,
+      "loss": 0.0739,
+      "step": 4040
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.1092,
+      "step": 4050
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 2.6071169208424112e-05,
+      "loss": 0.2856,
+      "step": 4060
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.599854756717502e-05,
+      "loss": 0.1629,
+      "step": 4070
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.2747,
+      "step": 4080
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": {
+        "accuracy": 0.9215686274509803
+      },
+      "eval_f1": {
+        "f1": 0.9215050761511078
+      },
+      "eval_loss": 0.3214215338230133,
+      "eval_precision": {
+        "precision": 0.923286387036387
+      },
+      "eval_recall": {
+        "recall": 0.921530375652716
+      },
+      "eval_runtime": 10.0587,
+      "eval_samples_per_second": 50.702,
+      "eval_steps_per_second": 25.351,
+      "step": 4080
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.5294065500042035e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-4080/training_args.bin b/tmp-checkpoint-4080/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-4080/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-4335/config.json b/tmp-checkpoint-4335/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-4335/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-4335/model.safetensors b/tmp-checkpoint-4335/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..894cd6c1958be4ceee9daa2a956cee2c49fda80e
--- /dev/null
+++ b/tmp-checkpoint-4335/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:636af36f52e1018e8eb5883246491ccb861144e69e6322e24bf963ab3888569c
+size 343233204
diff --git a/tmp-checkpoint-4335/optimizer.pt b/tmp-checkpoint-4335/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a8f817746827b5eb4abec9b7945b053a55a027bf
--- /dev/null
+++ b/tmp-checkpoint-4335/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25566a460fe530c4258c2d016cea6519b2dc2e20de8e9992cbcfe704dc46fcb4
+size 686587322
diff --git a/tmp-checkpoint-4335/preprocessor_config.json b/tmp-checkpoint-4335/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-4335/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-4335/rng_state.pth b/tmp-checkpoint-4335/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..5e4e63ea2e41c2e02311488d8fffb6fd54362fc9
--- /dev/null
+++ b/tmp-checkpoint-4335/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dba8ae74935f653d1c91731bc2d0eedc8d14b8182de7472328ee2a3725319b37
+size 14244
diff --git a/tmp-checkpoint-4335/scheduler.pt b/tmp-checkpoint-4335/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7c6fd07bca395883d64b6305a30cc429e86abac5
--- /dev/null
+++ b/tmp-checkpoint-4335/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4333457314b83f5dee5e39f80d2c9d7db7b4047a9d563b47a6e8c6dc0edeaffe
+size 1064
diff --git a/tmp-checkpoint-4335/trainer_state.json b/tmp-checkpoint-4335/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..53bf350157f39c9b023395462fc26a39bb3747f4
--- /dev/null
+++ b/tmp-checkpoint-4335/trainer_state.json
@@ -0,0 +1,2959 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 17.0,
+  "eval_steps": 500,
+  "global_step": 4335,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.1445170660856935e-05,
+      "loss": 0.1158,
+      "step": 3320
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.1086,
+      "step": 3330
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.129992737835875e-05,
+      "loss": 0.1612,
+      "step": 3340
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.122730573710966e-05,
+      "loss": 0.1272,
+      "step": 3350
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.115468409586057e-05,
+      "loss": 0.2715,
+      "step": 3360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.1082062454611474e-05,
+      "loss": 0.1432,
+      "step": 3370
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.100944081336238e-05,
+      "loss": 0.0573,
+      "step": 3380
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.093681917211329e-05,
+      "loss": 0.141,
+      "step": 3390
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.08641975308642e-05,
+      "loss": 0.156,
+      "step": 3400
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.0791575889615106e-05,
+      "loss": 0.1558,
+      "step": 3410
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.1606,
+      "step": 3420
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.064633260711692e-05,
+      "loss": 0.1677,
+      "step": 3430
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.057371096586783e-05,
+      "loss": 0.0562,
+      "step": 3440
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.0501089324618738e-05,
+      "loss": 0.1684,
+      "step": 3450
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.0428467683369643e-05,
+      "loss": 0.3589,
+      "step": 3460
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.0355846042120557e-05,
+      "loss": 0.2076,
+      "step": 3470
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.0283224400871462e-05,
+      "loss": 0.0773,
+      "step": 3480
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.021060275962237e-05,
+      "loss": 0.0424,
+      "step": 3490
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.0137981118373278e-05,
+      "loss": 0.1573,
+      "step": 3500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.0465,
+      "step": 3510
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.999273783587509e-05,
+      "loss": 0.0091,
+      "step": 3520
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.9920116194626e-05,
+      "loss": 0.2262,
+      "step": 3530
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.984749455337691e-05,
+      "loss": 0.0253,
+      "step": 3540
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.9774872912127817e-05,
+      "loss": 0.184,
+      "step": 3550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.9702251270878722e-05,
+      "loss": 0.1289,
+      "step": 3560
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.2017,
+      "step": 3570
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9274509803921569
+      },
+      "eval_f1": {
+        "f1": 0.9275152403725876
+      },
+      "eval_loss": 0.3324253261089325,
+      "eval_precision": {
+        "precision": 0.9309834626869261
+      },
+      "eval_recall": {
+        "recall": 0.9275085729872963
+      },
+      "eval_runtime": 10.2109,
+      "eval_samples_per_second": 49.947,
+      "eval_steps_per_second": 24.973,
+      "step": 3570
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 2.9557007988380538e-05,
+      "loss": 0.2156,
+      "step": 3580
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 2.948438634713145e-05,
+      "loss": 0.1672,
+      "step": 3590
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.0842,
+      "step": 3600
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 2.9339143064633262e-05,
+      "loss": 0.1529,
+      "step": 3610
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 2.926652142338417e-05,
+      "loss": 0.0904,
+      "step": 3620
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 2.9193899782135074e-05,
+      "loss": 0.1129,
+      "step": 3630
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.912127814088599e-05,
+      "loss": 0.0721,
+      "step": 3640
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 2.9048656499636893e-05,
+      "loss": 0.137,
+      "step": 3650
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 2.89760348583878e-05,
+      "loss": 0.1637,
+      "step": 3660
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.890341321713871e-05,
+      "loss": 0.1298,
+      "step": 3670
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 2.8830791575889614e-05,
+      "loss": 0.037,
+      "step": 3680
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.1013,
+      "step": 3690
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 2.8685548293391433e-05,
+      "loss": 0.1149,
+      "step": 3700
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 2.861292665214234e-05,
+      "loss": 0.1502,
+      "step": 3710
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 2.854030501089325e-05,
+      "loss": 0.1915,
+      "step": 3720
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 2.8467683369644154e-05,
+      "loss": 0.0854,
+      "step": 3730
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 2.839506172839506e-05,
+      "loss": 0.0395,
+      "step": 3740
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.832244008714597e-05,
+      "loss": 0.1681,
+      "step": 3750
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.824981844589688e-05,
+      "loss": 0.1236,
+      "step": 3760
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.817719680464779e-05,
+      "loss": 0.1138,
+      "step": 3770
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.1743,
+      "step": 3780
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.80319535221496e-05,
+      "loss": 0.0698,
+      "step": 3790
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.795933188090051e-05,
+      "loss": 0.1649,
+      "step": 3800
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.788671023965142e-05,
+      "loss": 0.0095,
+      "step": 3810
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.781408859840233e-05,
+      "loss": 0.017,
+      "step": 3820
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.9235294117647059
+      },
+      "eval_f1": {
+        "f1": 0.9236114478832926
+      },
+      "eval_loss": 0.319390207529068,
+      "eval_precision": {
+        "precision": 0.9261443200762086
+      },
+      "eval_recall": {
+        "recall": 0.9255596796820201
+      },
+      "eval_runtime": 10.1385,
+      "eval_samples_per_second": 50.303,
+      "eval_steps_per_second": 25.152,
+      "step": 3825
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 2.7741466957153233e-05,
+      "loss": 0.007,
+      "step": 3830
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.766884531590414e-05,
+      "loss": 0.1578,
+      "step": 3840
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 2.759622367465505e-05,
+      "loss": 0.0678,
+      "step": 3850
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 2.7523602033405954e-05,
+      "loss": 0.0545,
+      "step": 3860
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.1241,
+      "step": 3870
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.7378358750907773e-05,
+      "loss": 0.1044,
+      "step": 3880
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 2.730573710965868e-05,
+      "loss": 0.2882,
+      "step": 3890
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.7233115468409585e-05,
+      "loss": 0.1343,
+      "step": 3900
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.7160493827160493e-05,
+      "loss": 0.0315,
+      "step": 3910
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 2.70878721859114e-05,
+      "loss": 0.2315,
+      "step": 3920
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 2.7015250544662313e-05,
+      "loss": 0.0577,
+      "step": 3930
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 2.694262890341322e-05,
+      "loss": 0.1342,
+      "step": 3940
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.6870007262164125e-05,
+      "loss": 0.0996,
+      "step": 3950
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.1453,
+      "step": 3960
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 2.672476397966594e-05,
+      "loss": 0.0898,
+      "step": 3970
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.6652142338416852e-05,
+      "loss": 0.0078,
+      "step": 3980
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 2.657952069716776e-05,
+      "loss": 0.0463,
+      "step": 3990
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 2.6506899055918665e-05,
+      "loss": 0.1273,
+      "step": 4000
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.6434277414669573e-05,
+      "loss": 0.0209,
+      "step": 4010
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.636165577342048e-05,
+      "loss": 0.0307,
+      "step": 4020
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.6289034132171385e-05,
+      "loss": 0.1268,
+      "step": 4030
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.62164124909223e-05,
+      "loss": 0.0739,
+      "step": 4040
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.1092,
+      "step": 4050
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 2.6071169208424112e-05,
+      "loss": 0.2856,
+      "step": 4060
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.599854756717502e-05,
+      "loss": 0.1629,
+      "step": 4070
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.2747,
+      "step": 4080
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": {
+        "accuracy": 0.9215686274509803
+      },
+      "eval_f1": {
+        "f1": 0.9215050761511078
+      },
+      "eval_loss": 0.3214215338230133,
+      "eval_precision": {
+        "precision": 0.923286387036387
+      },
+      "eval_recall": {
+        "recall": 0.921530375652716
+      },
+      "eval_runtime": 10.0587,
+      "eval_samples_per_second": 50.702,
+      "eval_steps_per_second": 25.351,
+      "step": 4080
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 2.5853304284676833e-05,
+      "loss": 0.0519,
+      "step": 4090
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 2.5780682643427744e-05,
+      "loss": 0.1616,
+      "step": 4100
+    },
+    {
+      "epoch": 16.12,
+      "learning_rate": 2.5708061002178652e-05,
+      "loss": 0.061,
+      "step": 4110
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 2.563543936092956e-05,
+      "loss": 0.2667,
+      "step": 4120
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 2.5562817719680465e-05,
+      "loss": 0.1783,
+      "step": 4130
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.0504,
+      "step": 4140
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 2.5417574437182277e-05,
+      "loss": 0.1189,
+      "step": 4150
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 2.5344952795933192e-05,
+      "loss": 0.1827,
+      "step": 4160
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 2.5272331154684096e-05,
+      "loss": 0.0072,
+      "step": 4170
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 2.5199709513435004e-05,
+      "loss": 0.0668,
+      "step": 4180
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 2.5127087872185912e-05,
+      "loss": 0.0493,
+      "step": 4190
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 2.5054466230936817e-05,
+      "loss": 0.086,
+      "step": 4200
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 2.4981844589687728e-05,
+      "loss": 0.1581,
+      "step": 4210
+    },
+    {
+      "epoch": 16.55,
+      "learning_rate": 2.4909222948438636e-05,
+      "loss": 0.0961,
+      "step": 4220
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.0512,
+      "step": 4230
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 2.4763979665940452e-05,
+      "loss": 0.038,
+      "step": 4240
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 2.4691358024691357e-05,
+      "loss": 0.2144,
+      "step": 4250
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 2.4618736383442268e-05,
+      "loss": 0.1749,
+      "step": 4260
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 2.4546114742193176e-05,
+      "loss": 0.1474,
+      "step": 4270
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 2.4473493100944084e-05,
+      "loss": 0.0897,
+      "step": 4280
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 2.4400871459694992e-05,
+      "loss": 0.0866,
+      "step": 4290
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 2.4328249818445896e-05,
+      "loss": 0.0596,
+      "step": 4300
+    },
+    {
+      "epoch": 16.9,
+      "learning_rate": 2.4255628177196808e-05,
+      "loss": 0.0608,
+      "step": 4310
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.053,
+      "step": 4320
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 2.411038489469862e-05,
+      "loss": 0.0148,
+      "step": 4330
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.9008866348649631
+      },
+      "eval_loss": 0.4273129999637604,
+      "eval_precision": {
+        "precision": 0.9097813011730537
+      },
+      "eval_recall": {
+        "recall": 0.9019045670641415
+      },
+      "eval_runtime": 10.0655,
+      "eval_samples_per_second": 50.668,
+      "eval_steps_per_second": 25.334,
+      "step": 4335
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.687494459379466e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-4335/training_args.bin b/tmp-checkpoint-4335/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-4335/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728
diff --git a/tmp-checkpoint-4590/config.json b/tmp-checkpoint-4590/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..70f7c32ca3451c8948f77472e8ec5d1515313b52
--- /dev/null
+++ b/tmp-checkpoint-4590/config.json
@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "crack",
+    "1": "for review",
+    "2": "material deterioration",
+    "3": "repair",
+    "4": "water repair"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "crack": 0,
+    "for review": 1,
+    "material deterioration": 2,
+    "repair": 3,
+    "water repair": 4
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2"
+}
diff --git a/tmp-checkpoint-4590/model.safetensors b/tmp-checkpoint-4590/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..02b0ebd42a539c8320f98b1cf77921cac1509918
--- /dev/null
+++ b/tmp-checkpoint-4590/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45a8e9cf842600f235cd6591e45120a144ba7d680bf9ae4d9f92472afaa7a475
+size 343233204
diff --git a/tmp-checkpoint-4590/optimizer.pt b/tmp-checkpoint-4590/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d135fbf6947f529755e605c2aeb8559d92a7f8b9
--- /dev/null
+++ b/tmp-checkpoint-4590/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea1ab47ed578623f1ca49ceff2a1554f027cd3e336c6a37efad332363d28f3ae
+size 686587322
diff --git a/tmp-checkpoint-4590/preprocessor_config.json b/tmp-checkpoint-4590/preprocessor_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a3e96ca2d2afdba583423a8461f606b1c68992d
--- /dev/null
+++ b/tmp-checkpoint-4590/preprocessor_config.json
@@ -0,0 +1,22 @@
+{
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}
diff --git a/tmp-checkpoint-4590/rng_state.pth b/tmp-checkpoint-4590/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1e843977ed32bf076c1be81fbef7fa662d5b9523
--- /dev/null
+++ b/tmp-checkpoint-4590/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8424f06b193e7d5cc630f7d30356985ce19d3564fd086892459be224fe70486a
+size 14244
diff --git a/tmp-checkpoint-4590/scheduler.pt b/tmp-checkpoint-4590/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..30e33ffbc7307e0c51be1d5fac89466c32193e07
--- /dev/null
+++ b/tmp-checkpoint-4590/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2773cb56595281dee03029b62e25cc01dd9dd252f4f607d67eb6365e6835a00b
+size 1064
diff --git a/tmp-checkpoint-4590/trainer_state.json b/tmp-checkpoint-4590/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c988185683695ae1b838416470e0373ef4d76342
--- /dev/null
+++ b/tmp-checkpoint-4590/trainer_state.json
@@ -0,0 +1,3135 @@
+{
+  "best_metric": 0.2912794053554535,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-bridgedefectVIT\\checkpoint-3060",
+  "epoch": 18.0,
+  "eval_steps": 500,
+  "global_step": 4590,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04,
+      "learning_rate": 6.535947712418301e-07,
+      "loss": 1.6179,
+      "step": 10
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.3071895424836602e-06,
+      "loss": 1.6223,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.96078431372549e-06,
+      "loss": 1.6263,
+      "step": 30
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.6143790849673204e-06,
+      "loss": 1.601,
+      "step": 40
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 3.2679738562091506e-06,
+      "loss": 1.6074,
+      "step": 50
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.92156862745098e-06,
+      "loss": 1.5955,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5751633986928105e-06,
+      "loss": 1.5698,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.228758169934641e-06,
+      "loss": 1.5797,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 1.5531,
+      "step": 90
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 6.535947712418301e-06,
+      "loss": 1.5434,
+      "step": 100
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 7.1895424836601305e-06,
+      "loss": 1.5433,
+      "step": 110
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 7.84313725490196e-06,
+      "loss": 1.5329,
+      "step": 120
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 8.496732026143791e-06,
+      "loss": 1.4995,
+      "step": 130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.150326797385621e-06,
+      "loss": 1.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.803921568627451e-06,
+      "loss": 1.4863,
+      "step": 150
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.0457516339869281e-05,
+      "loss": 1.4209,
+      "step": 160
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.4354,
+      "step": 170
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.3572,
+      "step": 180
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.2418300653594772e-05,
+      "loss": 1.3714,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3071895424836602e-05,
+      "loss": 1.3406,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3725490196078432e-05,
+      "loss": 1.2574,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4379084967320261e-05,
+      "loss": 1.2965,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5032679738562091e-05,
+      "loss": 1.2021,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.568627450980392e-05,
+      "loss": 1.2072,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6339869281045753e-05,
+      "loss": 1.1457,
+      "step": 250
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": {
+        "accuracy": 0.6941176470588235
+      },
+      "eval_f1": {
+        "f1": 0.683792091280419
+      },
+      "eval_loss": 1.158827304840088,
+      "eval_precision": {
+        "precision": 0.7377494408375539
+      },
+      "eval_recall": {
+        "recall": 0.6941484490686618
+      },
+      "eval_runtime": 25.3262,
+      "eval_samples_per_second": 20.137,
+      "eval_steps_per_second": 10.069,
+      "step": 255
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.6993464052287582e-05,
+      "loss": 1.0924,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0007,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8300653594771242e-05,
+      "loss": 1.0054,
+      "step": 280
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.895424836601307e-05,
+      "loss": 0.9365,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9607843137254903e-05,
+      "loss": 0.8932,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.0261437908496734e-05,
+      "loss": 0.916,
+      "step": 310
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.0915032679738563e-05,
+      "loss": 0.8099,
+      "step": 320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.1568627450980395e-05,
+      "loss": 0.8211,
+      "step": 330
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.6861,
+      "step": 340
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.7379,
+      "step": 350
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.7349,
+      "step": 360
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.6907,
+      "step": 370
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.7376,
+      "step": 380
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.6827,
+      "step": 390
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.6969,
+      "step": 400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.6247,
+      "step": 410
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.5244,
+      "step": 420
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.7068,
+      "step": 430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.5623,
+      "step": 440
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.5931,
+      "step": 450
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.6641,
+      "step": 460
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.5438,
+      "step": 470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.5134,
+      "step": 480
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.4682,
+      "step": 490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.4762,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.4558,
+      "step": 510
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": {
+        "accuracy": 0.8470588235294118
+      },
+      "eval_f1": {
+        "f1": 0.8510200857474889
+      },
+      "eval_loss": 0.498457670211792,
+      "eval_precision": {
+        "precision": 0.8766865984788879
+      },
+      "eval_recall": {
+        "recall": 0.8454231938274491
+      },
+      "eval_runtime": 10.0517,
+      "eval_samples_per_second": 50.738,
+      "eval_steps_per_second": 25.369,
+      "step": 510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.4716,
+      "step": 520
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.5375,
+      "step": 530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.4725,
+      "step": 540
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.4804,
+      "step": 550
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.3504,
+      "step": 560
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.4049,
+      "step": 570
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.4309,
+      "step": 580
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.4938,
+      "step": 590
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.5427,
+      "step": 600
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.4719,
+      "step": 610
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.4627,
+      "step": 620
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.5277,
+      "step": 630
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.3109,
+      "step": 640
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.3799,
+      "step": 650
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.3515,
+      "step": 660
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.3722,
+      "step": 670
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.5337,
+      "step": 680
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.3263,
+      "step": 690
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.3703,
+      "step": 700
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.4498,
+      "step": 710
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.4109,
+      "step": 720
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.2953,
+      "step": 730
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.4033,
+      "step": 740
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.4186,
+      "step": 750
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3368,
+      "step": 760
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": {
+        "accuracy": 0.8705882352941177
+      },
+      "eval_f1": {
+        "f1": 0.8713991689505157
+      },
+      "eval_loss": 0.44290485978126526,
+      "eval_precision": {
+        "precision": 0.8884985994397759
+      },
+      "eval_recall": {
+        "recall": 0.8711626139817629
+      },
+      "eval_runtime": 9.9506,
+      "eval_samples_per_second": 51.253,
+      "eval_steps_per_second": 25.626,
+      "step": 765
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 4.9963689179375456e-05,
+      "loss": 0.3347,
+      "step": 770
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 4.9891067538126364e-05,
+      "loss": 0.3339,
+      "step": 780
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 4.981844589687727e-05,
+      "loss": 0.5573,
+      "step": 790
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 4.974582425562818e-05,
+      "loss": 0.5463,
+      "step": 800
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 4.967320261437909e-05,
+      "loss": 0.3453,
+      "step": 810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 4.9600580973129996e-05,
+      "loss": 0.3814,
+      "step": 820
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 4.9527959331880904e-05,
+      "loss": 0.4434,
+      "step": 830
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 4.945533769063181e-05,
+      "loss": 0.3675,
+      "step": 840
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 4.938271604938271e-05,
+      "loss": 0.5386,
+      "step": 850
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 4.931009440813363e-05,
+      "loss": 0.2692,
+      "step": 860
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9237472766884536e-05,
+      "loss": 0.4815,
+      "step": 870
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.9164851125635444e-05,
+      "loss": 0.4178,
+      "step": 880
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.909222948438635e-05,
+      "loss": 0.3618,
+      "step": 890
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.901960784313725e-05,
+      "loss": 0.6841,
+      "step": 900
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.894698620188817e-05,
+      "loss": 0.471,
+      "step": 910
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 4.8874364560639076e-05,
+      "loss": 0.313,
+      "step": 920
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8801742919389983e-05,
+      "loss": 0.2437,
+      "step": 930
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.8729121278140885e-05,
+      "loss": 0.4748,
+      "step": 940
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.865649963689179e-05,
+      "loss": 0.5519,
+      "step": 950
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 4.85838779956427e-05,
+      "loss": 0.3051,
+      "step": 960
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 4.8511256354393615e-05,
+      "loss": 0.3847,
+      "step": 970
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 4.843863471314452e-05,
+      "loss": 0.2454,
+      "step": 980
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 4.8366013071895424e-05,
+      "loss": 0.379,
+      "step": 990
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.829339143064633e-05,
+      "loss": 0.4485,
+      "step": 1000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.822076978939724e-05,
+      "loss": 0.2484,
+      "step": 1010
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.4402,
+      "step": 1020
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": {
+        "accuracy": 0.8784313725490196
+      },
+      "eval_f1": {
+        "f1": 0.880471094803992
+      },
+      "eval_loss": 0.3946979343891144,
+      "eval_precision": {
+        "precision": 0.8952890487430221
+      },
+      "eval_recall": {
+        "recall": 0.8776488582339645
+      },
+      "eval_runtime": 10.1771,
+      "eval_samples_per_second": 50.113,
+      "eval_steps_per_second": 25.056,
+      "step": 1020
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 4.807552650689906e-05,
+      "loss": 0.3494,
+      "step": 1030
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 4.8002904865649964e-05,
+      "loss": 0.3327,
+      "step": 1040
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 4.793028322440087e-05,
+      "loss": 0.2898,
+      "step": 1050
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 4.785766158315178e-05,
+      "loss": 0.3965,
+      "step": 1060
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 4.778503994190269e-05,
+      "loss": 0.3607,
+      "step": 1070
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 4.77124183006536e-05,
+      "loss": 0.3714,
+      "step": 1080
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 4.7639796659404504e-05,
+      "loss": 0.1796,
+      "step": 1090
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 4.756717501815541e-05,
+      "loss": 0.1874,
+      "step": 1100
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 4.749455337690632e-05,
+      "loss": 0.3462,
+      "step": 1110
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 4.742193173565723e-05,
+      "loss": 0.2082,
+      "step": 1120
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 4.7349310094408136e-05,
+      "loss": 0.4531,
+      "step": 1130
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 4.7276688453159044e-05,
+      "loss": 0.3951,
+      "step": 1140
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 4.720406681190995e-05,
+      "loss": 0.2337,
+      "step": 1150
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.713144517066086e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 4.705882352941177e-05,
+      "loss": 0.373,
+      "step": 1170
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 4.6986201888162675e-05,
+      "loss": 0.1484,
+      "step": 1180
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 4.691358024691358e-05,
+      "loss": 0.2568,
+      "step": 1190
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 4.684095860566449e-05,
+      "loss": 0.3067,
+      "step": 1200
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 4.67683369644154e-05,
+      "loss": 0.2007,
+      "step": 1210
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 4.669571532316631e-05,
+      "loss": 0.2493,
+      "step": 1220
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 4.6623093681917215e-05,
+      "loss": 0.2717,
+      "step": 1230
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.6550472040668116e-05,
+      "loss": 0.4134,
+      "step": 1240
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 4.647785039941903e-05,
+      "loss": 0.2417,
+      "step": 1250
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 4.640522875816994e-05,
+      "loss": 0.2573,
+      "step": 1260
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 4.633260711692085e-05,
+      "loss": 0.3937,
+      "step": 1270
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": {
+        "accuracy": 0.8725490196078431
+      },
+      "eval_f1": {
+        "f1": 0.8703780158419334
+      },
+      "eval_loss": 0.4377364218235016,
+      "eval_precision": {
+        "precision": 0.8900760323128502
+      },
+      "eval_recall": {
+        "recall": 0.8691138648585458
+      },
+      "eval_runtime": 10.1762,
+      "eval_samples_per_second": 50.117,
+      "eval_steps_per_second": 25.058,
+      "step": 1275
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 4.6259985475671755e-05,
+      "loss": 0.165,
+      "step": 1280
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 4.6187363834422656e-05,
+      "loss": 0.1452,
+      "step": 1290
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 4.6114742193173564e-05,
+      "loss": 0.1548,
+      "step": 1300
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 4.604212055192448e-05,
+      "loss": 0.4443,
+      "step": 1310
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 4.5969498910675387e-05,
+      "loss": 0.4322,
+      "step": 1320
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 4.5896877269426294e-05,
+      "loss": 0.2355,
+      "step": 1330
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 4.5824255628177196e-05,
+      "loss": 0.1949,
+      "step": 1340
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 4.5751633986928104e-05,
+      "loss": 0.1959,
+      "step": 1350
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 4.567901234567901e-05,
+      "loss": 0.3841,
+      "step": 1360
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 4.5606390704429926e-05,
+      "loss": 0.1401,
+      "step": 1370
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 4.5533769063180834e-05,
+      "loss": 0.1148,
+      "step": 1380
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 4.5461147421931735e-05,
+      "loss": 0.234,
+      "step": 1390
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 4.538852578068264e-05,
+      "loss": 0.3899,
+      "step": 1400
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 4.531590413943355e-05,
+      "loss": 0.4873,
+      "step": 1410
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 4.5243282498184466e-05,
+      "loss": 0.2483,
+      "step": 1420
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 4.5170660856935374e-05,
+      "loss": 0.2654,
+      "step": 1430
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 4.5098039215686275e-05,
+      "loss": 0.2035,
+      "step": 1440
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 4.502541757443718e-05,
+      "loss": 0.3744,
+      "step": 1450
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 4.495279593318809e-05,
+      "loss": 0.2139,
+      "step": 1460
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 4.4880174291939e-05,
+      "loss": 0.2212,
+      "step": 1470
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 4.480755265068991e-05,
+      "loss": 0.376,
+      "step": 1480
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 4.4734931009440815e-05,
+      "loss": 0.2399,
+      "step": 1490
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 4.466230936819172e-05,
+      "loss": 0.3534,
+      "step": 1500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 4.458968772694263e-05,
+      "loss": 0.3667,
+      "step": 1510
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 4.451706608569354e-05,
+      "loss": 0.496,
+      "step": 1520
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.4042,
+      "step": 1530
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8859196308355288
+      },
+      "eval_loss": 0.39891302585601807,
+      "eval_precision": {
+        "precision": 0.8972093144987638
+      },
+      "eval_recall": {
+        "recall": 0.8834955381497934
+      },
+      "eval_runtime": 10.1174,
+      "eval_samples_per_second": 50.408,
+      "eval_steps_per_second": 25.204,
+      "step": 1530
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 4.4371822803195355e-05,
+      "loss": 0.2637,
+      "step": 1540
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 4.429920116194626e-05,
+      "loss": 0.2821,
+      "step": 1550
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 4.422657952069717e-05,
+      "loss": 0.1201,
+      "step": 1560
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.415395787944808e-05,
+      "loss": 0.262,
+      "step": 1570
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 4.4081336238198986e-05,
+      "loss": 0.2064,
+      "step": 1580
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 4.400871459694989e-05,
+      "loss": 0.1815,
+      "step": 1590
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 4.39360929557008e-05,
+      "loss": 0.1959,
+      "step": 1600
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 4.386347131445171e-05,
+      "loss": 0.1959,
+      "step": 1610
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 4.379084967320262e-05,
+      "loss": 0.2398,
+      "step": 1620
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 4.3718228031953526e-05,
+      "loss": 0.2038,
+      "step": 1630
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 4.364560639070443e-05,
+      "loss": 0.2004,
+      "step": 1640
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 4.357298474945534e-05,
+      "loss": 0.0377,
+      "step": 1650
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 4.350036310820625e-05,
+      "loss": 0.2022,
+      "step": 1660
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 4.342774146695716e-05,
+      "loss": 0.4403,
+      "step": 1670
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 4.3355119825708066e-05,
+      "loss": 0.2577,
+      "step": 1680
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 4.328249818445897e-05,
+      "loss": 0.3312,
+      "step": 1690
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 4.3209876543209875e-05,
+      "loss": 0.307,
+      "step": 1700
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 4.313725490196079e-05,
+      "loss": 0.0961,
+      "step": 1710
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 4.30646332607117e-05,
+      "loss": 0.3266,
+      "step": 1720
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.2992011619462605e-05,
+      "loss": 0.2601,
+      "step": 1730
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.291938997821351e-05,
+      "loss": 0.2382,
+      "step": 1740
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.2846768336964415e-05,
+      "loss": 0.1469,
+      "step": 1750
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.277414669571532e-05,
+      "loss": 0.2158,
+      "step": 1760
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.270152505446624e-05,
+      "loss": 0.4392,
+      "step": 1770
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.262890341321714e-05,
+      "loss": 0.4321,
+      "step": 1780
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": {
+        "accuracy": 0.8901960784313725
+      },
+      "eval_f1": {
+        "f1": 0.8896787277785286
+      },
+      "eval_loss": 0.36420056223869324,
+      "eval_precision": {
+        "precision": 0.9002265766283616
+      },
+      "eval_recall": {
+        "recall": 0.8878911425453978
+      },
+      "eval_runtime": 10.2347,
+      "eval_samples_per_second": 49.83,
+      "eval_steps_per_second": 24.915,
+      "step": 1785
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.2556281771968046e-05,
+      "loss": 0.5162,
+      "step": 1790
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2483660130718954e-05,
+      "loss": 0.2491,
+      "step": 1800
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.241103848946986e-05,
+      "loss": 0.2267,
+      "step": 1810
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.233841684822078e-05,
+      "loss": 0.2395,
+      "step": 1820
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 4.226579520697168e-05,
+      "loss": 0.3007,
+      "step": 1830
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 4.2193173565722586e-05,
+      "loss": 0.1316,
+      "step": 1840
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 4.2120551924473494e-05,
+      "loss": 0.1592,
+      "step": 1850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 4.20479302832244e-05,
+      "loss": 0.3467,
+      "step": 1860
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 4.197530864197531e-05,
+      "loss": 0.4195,
+      "step": 1870
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 4.190268700072622e-05,
+      "loss": 0.3561,
+      "step": 1880
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 4.1830065359477126e-05,
+      "loss": 0.1572,
+      "step": 1890
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 4.1757443718228034e-05,
+      "loss": 0.2104,
+      "step": 1900
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 4.168482207697894e-05,
+      "loss": 0.3536,
+      "step": 1910
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 4.161220043572985e-05,
+      "loss": 0.345,
+      "step": 1920
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 4.153957879448076e-05,
+      "loss": 0.1083,
+      "step": 1930
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 4.1466957153231666e-05,
+      "loss": 0.1895,
+      "step": 1940
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 4.1394335511982573e-05,
+      "loss": 0.3232,
+      "step": 1950
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 4.132171387073348e-05,
+      "loss": 0.2939,
+      "step": 1960
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 4.124909222948439e-05,
+      "loss": 0.0918,
+      "step": 1970
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 4.11764705882353e-05,
+      "loss": 0.1722,
+      "step": 1980
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 4.1103848946986205e-05,
+      "loss": 0.3229,
+      "step": 1990
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 4.103122730573711e-05,
+      "loss": 0.1501,
+      "step": 2000
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 4.095860566448802e-05,
+      "loss": 0.1468,
+      "step": 2010
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 4.088598402323893e-05,
+      "loss": 0.1665,
+      "step": 2020
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 4.081336238198984e-05,
+      "loss": 0.1324,
+      "step": 2030
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2618,
+      "step": 2040
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": {
+        "accuracy": 0.8960784313725491
+      },
+      "eval_f1": {
+        "f1": 0.8950352545633242
+      },
+      "eval_loss": 0.40923169255256653,
+      "eval_precision": {
+        "precision": 0.9033601037449228
+      },
+      "eval_recall": {
+        "recall": 0.8953584093211753
+      },
+      "eval_runtime": 10.3235,
+      "eval_samples_per_second": 49.402,
+      "eval_steps_per_second": 24.701,
+      "step": 2040
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 4.066811909949165e-05,
+      "loss": 0.1887,
+      "step": 2050
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 4.059549745824256e-05,
+      "loss": 0.4153,
+      "step": 2060
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 4.052287581699347e-05,
+      "loss": 0.2771,
+      "step": 2070
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 4.045025417574437e-05,
+      "loss": 0.3316,
+      "step": 2080
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 4.037763253449528e-05,
+      "loss": 0.1125,
+      "step": 2090
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 4.0305010893246186e-05,
+      "loss": 0.3016,
+      "step": 2100
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 4.02323892519971e-05,
+      "loss": 0.2458,
+      "step": 2110
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 4.015976761074801e-05,
+      "loss": 0.179,
+      "step": 2120
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 4.008714596949891e-05,
+      "loss": 0.224,
+      "step": 2130
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 4.001452432824982e-05,
+      "loss": 0.0803,
+      "step": 2140
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 3.9941902687000726e-05,
+      "loss": 0.1121,
+      "step": 2150
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 3.986928104575164e-05,
+      "loss": 0.3751,
+      "step": 2160
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 3.979665940450255e-05,
+      "loss": 0.1354,
+      "step": 2170
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 3.972403776325345e-05,
+      "loss": 0.3105,
+      "step": 2180
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 3.965141612200436e-05,
+      "loss": 0.1937,
+      "step": 2190
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 3.9578794480755265e-05,
+      "loss": 0.3028,
+      "step": 2200
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 3.950617283950617e-05,
+      "loss": 0.1851,
+      "step": 2210
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 3.943355119825709e-05,
+      "loss": 0.074,
+      "step": 2220
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 3.936092955700799e-05,
+      "loss": 0.2098,
+      "step": 2230
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 3.92883079157589e-05,
+      "loss": 0.3167,
+      "step": 2240
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 3.9215686274509805e-05,
+      "loss": 0.2527,
+      "step": 2250
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 3.914306463326071e-05,
+      "loss": 0.1521,
+      "step": 2260
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 3.907044299201162e-05,
+      "loss": 0.1393,
+      "step": 2270
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 3.899782135076253e-05,
+      "loss": 0.1155,
+      "step": 2280
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 3.892519970951344e-05,
+      "loss": 0.3376,
+      "step": 2290
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": {
+        "accuracy": 0.8941176470588236
+      },
+      "eval_f1": {
+        "f1": 0.8933906521569781
+      },
+      "eval_loss": 0.4055422246456146,
+      "eval_precision": {
+        "precision": 0.9014239716079142
+      },
+      "eval_recall": {
+        "recall": 0.8936214051905541
+      },
+      "eval_runtime": 10.1735,
+      "eval_samples_per_second": 50.13,
+      "eval_steps_per_second": 25.065,
+      "step": 2295
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 3.8852578068264345e-05,
+      "loss": 0.1807,
+      "step": 2300
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 3.877995642701525e-05,
+      "loss": 0.1127,
+      "step": 2310
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 3.870733478576616e-05,
+      "loss": 0.0503,
+      "step": 2320
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.863471314451707e-05,
+      "loss": 0.0748,
+      "step": 2330
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.8562091503267977e-05,
+      "loss": 0.2657,
+      "step": 2340
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.8489469862018884e-05,
+      "loss": 0.1785,
+      "step": 2350
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.841684822076979e-05,
+      "loss": 0.0808,
+      "step": 2360
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.83442265795207e-05,
+      "loss": 0.3348,
+      "step": 2370
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.259,
+      "step": 2380
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.8198983297022516e-05,
+      "loss": 0.1707,
+      "step": 2390
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 3.8126361655773424e-05,
+      "loss": 0.0915,
+      "step": 2400
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 3.805374001452433e-05,
+      "loss": 0.0987,
+      "step": 2410
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 3.798111837327524e-05,
+      "loss": 0.0906,
+      "step": 2420
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 3.790849673202614e-05,
+      "loss": 0.1955,
+      "step": 2430
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 3.783587509077705e-05,
+      "loss": 0.1936,
+      "step": 2440
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 3.7763253449527964e-05,
+      "loss": 0.1479,
+      "step": 2450
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 3.769063180827887e-05,
+      "loss": 0.1662,
+      "step": 2460
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 3.761801016702978e-05,
+      "loss": 0.0815,
+      "step": 2470
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.754538852578068e-05,
+      "loss": 0.1128,
+      "step": 2480
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.747276688453159e-05,
+      "loss": 0.0891,
+      "step": 2490
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 3.74001452432825e-05,
+      "loss": 0.0853,
+      "step": 2500
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 3.732752360203341e-05,
+      "loss": 0.2403,
+      "step": 2510
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 3.725490196078432e-05,
+      "loss": 0.2393,
+      "step": 2520
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.718228031953522e-05,
+      "loss": 0.113,
+      "step": 2530
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 3.710965867828613e-05,
+      "loss": 0.1374,
+      "step": 2540
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.2592,
+      "step": 2550
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": {
+        "accuracy": 0.8627450980392157
+      },
+      "eval_f1": {
+        "f1": 0.8639402375974159
+      },
+      "eval_loss": 0.49467796087265015,
+      "eval_precision": {
+        "precision": 0.8837020302136581
+      },
+      "eval_recall": {
+        "recall": 0.8620484568622867
+      },
+      "eval_runtime": 10.09,
+      "eval_samples_per_second": 50.545,
+      "eval_steps_per_second": 25.273,
+      "step": 2550
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 3.696441539578795e-05,
+      "loss": 0.1971,
+      "step": 2560
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 3.689179375453886e-05,
+      "loss": 0.1467,
+      "step": 2570
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 3.681917211328976e-05,
+      "loss": 0.2304,
+      "step": 2580
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 3.674655047204067e-05,
+      "loss": 0.1111,
+      "step": 2590
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 3.6673928830791576e-05,
+      "loss": 0.2193,
+      "step": 2600
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 3.6601307189542484e-05,
+      "loss": 0.1701,
+      "step": 2610
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 3.652868554829339e-05,
+      "loss": 0.0251,
+      "step": 2620
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 3.64560639070443e-05,
+      "loss": 0.0138,
+      "step": 2630
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 3.638344226579521e-05,
+      "loss": 0.191,
+      "step": 2640
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 3.6310820624546116e-05,
+      "loss": 0.1112,
+      "step": 2650
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 3.6238198983297024e-05,
+      "loss": 0.1115,
+      "step": 2660
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 3.616557734204793e-05,
+      "loss": 0.1842,
+      "step": 2670
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 3.609295570079884e-05,
+      "loss": 0.1397,
+      "step": 2680
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 3.602033405954975e-05,
+      "loss": 0.1044,
+      "step": 2690
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 3.5947712418300656e-05,
+      "loss": 0.1015,
+      "step": 2700
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 3.5875090777051564e-05,
+      "loss": 0.2076,
+      "step": 2710
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 3.580246913580247e-05,
+      "loss": 0.1405,
+      "step": 2720
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 3.572984749455338e-05,
+      "loss": 0.154,
+      "step": 2730
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 3.565722585330429e-05,
+      "loss": 0.2566,
+      "step": 2740
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 3.5584604212055195e-05,
+      "loss": 0.1589,
+      "step": 2750
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 3.55119825708061e-05,
+      "loss": 0.2138,
+      "step": 2760
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 3.543936092955701e-05,
+      "loss": 0.1734,
+      "step": 2770
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 3.536673928830791e-05,
+      "loss": 0.0846,
+      "step": 2780
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 3.529411764705883e-05,
+      "loss": 0.2093,
+      "step": 2790
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 3.5221496005809735e-05,
+      "loss": 0.363,
+      "step": 2800
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": {
+        "accuracy": 0.8862745098039215
+      },
+      "eval_f1": {
+        "f1": 0.8871534026726227
+      },
+      "eval_loss": 0.45857715606689453,
+      "eval_precision": {
+        "precision": 0.89776885319054
+      },
+      "eval_recall": {
+        "recall": 0.8850722858701582
+      },
+      "eval_runtime": 9.9484,
+      "eval_samples_per_second": 51.265,
+      "eval_steps_per_second": 25.632,
+      "step": 2805
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 3.514887436456064e-05,
+      "loss": 0.161,
+      "step": 2810
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.507625272331155e-05,
+      "loss": 0.0835,
+      "step": 2820
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.500363108206245e-05,
+      "loss": 0.1329,
+      "step": 2830
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.493100944081336e-05,
+      "loss": 0.0135,
+      "step": 2840
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4858387799564275e-05,
+      "loss": 0.0124,
+      "step": 2850
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.478576615831518e-05,
+      "loss": 0.2865,
+      "step": 2860
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.471314451706609e-05,
+      "loss": 0.2016,
+      "step": 2870
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 3.464052287581699e-05,
+      "loss": 0.1214,
+      "step": 2880
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 3.45679012345679e-05,
+      "loss": 0.1795,
+      "step": 2890
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 3.4495279593318815e-05,
+      "loss": 0.028,
+      "step": 2900
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 3.442265795206972e-05,
+      "loss": 0.3366,
+      "step": 2910
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 3.4350036310820624e-05,
+      "loss": 0.161,
+      "step": 2920
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 3.427741466957153e-05,
+      "loss": 0.1429,
+      "step": 2930
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 3.420479302832244e-05,
+      "loss": 0.1995,
+      "step": 2940
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 3.413217138707335e-05,
+      "loss": 0.0369,
+      "step": 2950
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 3.405954974582426e-05,
+      "loss": 0.1814,
+      "step": 2960
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 3.3986928104575163e-05,
+      "loss": 0.1466,
+      "step": 2970
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 3.391430646332607e-05,
+      "loss": 0.0435,
+      "step": 2980
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 3.384168482207698e-05,
+      "loss": 0.0945,
+      "step": 2990
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 3.376906318082789e-05,
+      "loss": 0.3593,
+      "step": 3000
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 3.3696441539578795e-05,
+      "loss": 0.1862,
+      "step": 3010
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 3.36238198983297e-05,
+      "loss": 0.1031,
+      "step": 3020
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 3.355119825708061e-05,
+      "loss": 0.1794,
+      "step": 3030
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.347857661583152e-05,
+      "loss": 0.2495,
+      "step": 3040
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 3.340595497458243e-05,
+      "loss": 0.168,
+      "step": 3050
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.278,
+      "step": 3060
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": {
+        "accuracy": 0.9196078431372549
+      },
+      "eval_f1": {
+        "f1": 0.9203616179126926
+      },
+      "eval_loss": 0.2912794053554535,
+      "eval_precision": {
+        "precision": 0.9237222880261108
+      },
+      "eval_recall": {
+        "recall": 0.920318466993999
+      },
+      "eval_runtime": 10.2193,
+      "eval_samples_per_second": 49.905,
+      "eval_steps_per_second": 24.953,
+      "step": 3060
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 3.326071169208424e-05,
+      "loss": 0.0654,
+      "step": 3070
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 3.318809005083515e-05,
+      "loss": 0.1827,
+      "step": 3080
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.311546840958606e-05,
+      "loss": 0.1608,
+      "step": 3090
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.304284676833697e-05,
+      "loss": 0.1133,
+      "step": 3100
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.2970225127087875e-05,
+      "loss": 0.1324,
+      "step": 3110
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.289760348583878e-05,
+      "loss": 0.233,
+      "step": 3120
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.282498184458969e-05,
+      "loss": 0.1262,
+      "step": 3130
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.27523602033406e-05,
+      "loss": 0.1989,
+      "step": 3140
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.2679738562091506e-05,
+      "loss": 0.0913,
+      "step": 3150
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.2607116920842414e-05,
+      "loss": 0.14,
+      "step": 3160
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.2534495279593316e-05,
+      "loss": 0.1296,
+      "step": 3170
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.2461873638344223e-05,
+      "loss": 0.137,
+      "step": 3180
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.238925199709514e-05,
+      "loss": 0.2531,
+      "step": 3190
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.2316630355846046e-05,
+      "loss": 0.1718,
+      "step": 3200
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.2244008714596954e-05,
+      "loss": 0.0158,
+      "step": 3210
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.2171387073347855e-05,
+      "loss": 0.2503,
+      "step": 3220
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.209876543209876e-05,
+      "loss": 0.0101,
+      "step": 3230
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.202614379084967e-05,
+      "loss": 0.1985,
+      "step": 3240
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.1953522149600586e-05,
+      "loss": 0.2394,
+      "step": 3250
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.1880900508351494e-05,
+      "loss": 0.183,
+      "step": 3260
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.1808278867102395e-05,
+      "loss": 0.1489,
+      "step": 3270
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.17356572258533e-05,
+      "loss": 0.0722,
+      "step": 3280
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.166303558460421e-05,
+      "loss": 0.0889,
+      "step": 3290
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.1590413943355126e-05,
+      "loss": 0.1194,
+      "step": 3300
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.1517792302106034e-05,
+      "loss": 0.1436,
+      "step": 3310
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.8997771222806092
+      },
+      "eval_loss": 0.4294499456882477,
+      "eval_precision": {
+        "precision": 0.9095799880746632
+      },
+      "eval_recall": {
+        "recall": 0.901510989010989
+      },
+      "eval_runtime": 10.2735,
+      "eval_samples_per_second": 49.642,
+      "eval_steps_per_second": 24.821,
+      "step": 3315
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.1445170660856935e-05,
+      "loss": 0.1158,
+      "step": 3320
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.137254901960784e-05,
+      "loss": 0.1086,
+      "step": 3330
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.129992737835875e-05,
+      "loss": 0.1612,
+      "step": 3340
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.122730573710966e-05,
+      "loss": 0.1272,
+      "step": 3350
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.115468409586057e-05,
+      "loss": 0.2715,
+      "step": 3360
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.1082062454611474e-05,
+      "loss": 0.1432,
+      "step": 3370
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.100944081336238e-05,
+      "loss": 0.0573,
+      "step": 3380
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.093681917211329e-05,
+      "loss": 0.141,
+      "step": 3390
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.08641975308642e-05,
+      "loss": 0.156,
+      "step": 3400
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.0791575889615106e-05,
+      "loss": 0.1558,
+      "step": 3410
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.0718954248366014e-05,
+      "loss": 0.1606,
+      "step": 3420
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.064633260711692e-05,
+      "loss": 0.1677,
+      "step": 3430
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.057371096586783e-05,
+      "loss": 0.0562,
+      "step": 3440
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.0501089324618738e-05,
+      "loss": 0.1684,
+      "step": 3450
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.0428467683369643e-05,
+      "loss": 0.3589,
+      "step": 3460
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.0355846042120557e-05,
+      "loss": 0.2076,
+      "step": 3470
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.0283224400871462e-05,
+      "loss": 0.0773,
+      "step": 3480
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.021060275962237e-05,
+      "loss": 0.0424,
+      "step": 3490
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.0137981118373278e-05,
+      "loss": 0.1573,
+      "step": 3500
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.0065359477124182e-05,
+      "loss": 0.0465,
+      "step": 3510
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 2.999273783587509e-05,
+      "loss": 0.0091,
+      "step": 3520
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 2.9920116194626e-05,
+      "loss": 0.2262,
+      "step": 3530
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 2.984749455337691e-05,
+      "loss": 0.0253,
+      "step": 3540
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 2.9774872912127817e-05,
+      "loss": 0.184,
+      "step": 3550
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 2.9702251270878722e-05,
+      "loss": 0.1289,
+      "step": 3560
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.2017,
+      "step": 3570
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": {
+        "accuracy": 0.9274509803921569
+      },
+      "eval_f1": {
+        "f1": 0.9275152403725876
+      },
+      "eval_loss": 0.3324253261089325,
+      "eval_precision": {
+        "precision": 0.9309834626869261
+      },
+      "eval_recall": {
+        "recall": 0.9275085729872963
+      },
+      "eval_runtime": 10.2109,
+      "eval_samples_per_second": 49.947,
+      "eval_steps_per_second": 24.973,
+      "step": 3570
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 2.9557007988380538e-05,
+      "loss": 0.2156,
+      "step": 3580
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 2.948438634713145e-05,
+      "loss": 0.1672,
+      "step": 3590
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 0.0842,
+      "step": 3600
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 2.9339143064633262e-05,
+      "loss": 0.1529,
+      "step": 3610
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 2.926652142338417e-05,
+      "loss": 0.0904,
+      "step": 3620
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 2.9193899782135074e-05,
+      "loss": 0.1129,
+      "step": 3630
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.912127814088599e-05,
+      "loss": 0.0721,
+      "step": 3640
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 2.9048656499636893e-05,
+      "loss": 0.137,
+      "step": 3650
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 2.89760348583878e-05,
+      "loss": 0.1637,
+      "step": 3660
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.890341321713871e-05,
+      "loss": 0.1298,
+      "step": 3670
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 2.8830791575889614e-05,
+      "loss": 0.037,
+      "step": 3680
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 2.8758169934640522e-05,
+      "loss": 0.1013,
+      "step": 3690
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 2.8685548293391433e-05,
+      "loss": 0.1149,
+      "step": 3700
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 2.861292665214234e-05,
+      "loss": 0.1502,
+      "step": 3710
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 2.854030501089325e-05,
+      "loss": 0.1915,
+      "step": 3720
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 2.8467683369644154e-05,
+      "loss": 0.0854,
+      "step": 3730
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 2.839506172839506e-05,
+      "loss": 0.0395,
+      "step": 3740
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.832244008714597e-05,
+      "loss": 0.1681,
+      "step": 3750
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.824981844589688e-05,
+      "loss": 0.1236,
+      "step": 3760
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.817719680464779e-05,
+      "loss": 0.1138,
+      "step": 3770
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 2.8104575163398693e-05,
+      "loss": 0.1743,
+      "step": 3780
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.80319535221496e-05,
+      "loss": 0.0698,
+      "step": 3790
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.795933188090051e-05,
+      "loss": 0.1649,
+      "step": 3800
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.788671023965142e-05,
+      "loss": 0.0095,
+      "step": 3810
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.781408859840233e-05,
+      "loss": 0.017,
+      "step": 3820
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": {
+        "accuracy": 0.9235294117647059
+      },
+      "eval_f1": {
+        "f1": 0.9236114478832926
+      },
+      "eval_loss": 0.319390207529068,
+      "eval_precision": {
+        "precision": 0.9261443200762086
+      },
+      "eval_recall": {
+        "recall": 0.9255596796820201
+      },
+      "eval_runtime": 10.1385,
+      "eval_samples_per_second": 50.303,
+      "eval_steps_per_second": 25.152,
+      "step": 3825
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 2.7741466957153233e-05,
+      "loss": 0.007,
+      "step": 3830
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.766884531590414e-05,
+      "loss": 0.1578,
+      "step": 3840
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 2.759622367465505e-05,
+      "loss": 0.0678,
+      "step": 3850
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 2.7523602033405954e-05,
+      "loss": 0.0545,
+      "step": 3860
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.7450980392156865e-05,
+      "loss": 0.1241,
+      "step": 3870
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.7378358750907773e-05,
+      "loss": 0.1044,
+      "step": 3880
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 2.730573710965868e-05,
+      "loss": 0.2882,
+      "step": 3890
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.7233115468409585e-05,
+      "loss": 0.1343,
+      "step": 3900
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.7160493827160493e-05,
+      "loss": 0.0315,
+      "step": 3910
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 2.70878721859114e-05,
+      "loss": 0.2315,
+      "step": 3920
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 2.7015250544662313e-05,
+      "loss": 0.0577,
+      "step": 3930
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 2.694262890341322e-05,
+      "loss": 0.1342,
+      "step": 3940
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.6870007262164125e-05,
+      "loss": 0.0996,
+      "step": 3950
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.6797385620915033e-05,
+      "loss": 0.1453,
+      "step": 3960
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 2.672476397966594e-05,
+      "loss": 0.0898,
+      "step": 3970
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.6652142338416852e-05,
+      "loss": 0.0078,
+      "step": 3980
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 2.657952069716776e-05,
+      "loss": 0.0463,
+      "step": 3990
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 2.6506899055918665e-05,
+      "loss": 0.1273,
+      "step": 4000
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.6434277414669573e-05,
+      "loss": 0.0209,
+      "step": 4010
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.636165577342048e-05,
+      "loss": 0.0307,
+      "step": 4020
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.6289034132171385e-05,
+      "loss": 0.1268,
+      "step": 4030
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.62164124909223e-05,
+      "loss": 0.0739,
+      "step": 4040
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 2.6143790849673204e-05,
+      "loss": 0.1092,
+      "step": 4050
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 2.6071169208424112e-05,
+      "loss": 0.2856,
+      "step": 4060
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.599854756717502e-05,
+      "loss": 0.1629,
+      "step": 4070
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.2747,
+      "step": 4080
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": {
+        "accuracy": 0.9215686274509803
+      },
+      "eval_f1": {
+        "f1": 0.9215050761511078
+      },
+      "eval_loss": 0.3214215338230133,
+      "eval_precision": {
+        "precision": 0.923286387036387
+      },
+      "eval_recall": {
+        "recall": 0.921530375652716
+      },
+      "eval_runtime": 10.0587,
+      "eval_samples_per_second": 50.702,
+      "eval_steps_per_second": 25.351,
+      "step": 4080
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 2.5853304284676833e-05,
+      "loss": 0.0519,
+      "step": 4090
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 2.5780682643427744e-05,
+      "loss": 0.1616,
+      "step": 4100
+    },
+    {
+      "epoch": 16.12,
+      "learning_rate": 2.5708061002178652e-05,
+      "loss": 0.061,
+      "step": 4110
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 2.563543936092956e-05,
+      "loss": 0.2667,
+      "step": 4120
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 2.5562817719680465e-05,
+      "loss": 0.1783,
+      "step": 4130
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 2.5490196078431373e-05,
+      "loss": 0.0504,
+      "step": 4140
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 2.5417574437182277e-05,
+      "loss": 0.1189,
+      "step": 4150
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 2.5344952795933192e-05,
+      "loss": 0.1827,
+      "step": 4160
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 2.5272331154684096e-05,
+      "loss": 0.0072,
+      "step": 4170
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 2.5199709513435004e-05,
+      "loss": 0.0668,
+      "step": 4180
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 2.5127087872185912e-05,
+      "loss": 0.0493,
+      "step": 4190
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 2.5054466230936817e-05,
+      "loss": 0.086,
+      "step": 4200
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 2.4981844589687728e-05,
+      "loss": 0.1581,
+      "step": 4210
+    },
+    {
+      "epoch": 16.55,
+      "learning_rate": 2.4909222948438636e-05,
+      "loss": 0.0961,
+      "step": 4220
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 2.4836601307189544e-05,
+      "loss": 0.0512,
+      "step": 4230
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 2.4763979665940452e-05,
+      "loss": 0.038,
+      "step": 4240
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 2.4691358024691357e-05,
+      "loss": 0.2144,
+      "step": 4250
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 2.4618736383442268e-05,
+      "loss": 0.1749,
+      "step": 4260
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 2.4546114742193176e-05,
+      "loss": 0.1474,
+      "step": 4270
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 2.4473493100944084e-05,
+      "loss": 0.0897,
+      "step": 4280
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 2.4400871459694992e-05,
+      "loss": 0.0866,
+      "step": 4290
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 2.4328249818445896e-05,
+      "loss": 0.0596,
+      "step": 4300
+    },
+    {
+      "epoch": 16.9,
+      "learning_rate": 2.4255628177196808e-05,
+      "loss": 0.0608,
+      "step": 4310
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 2.4183006535947712e-05,
+      "loss": 0.053,
+      "step": 4320
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 2.411038489469862e-05,
+      "loss": 0.0148,
+      "step": 4330
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": {
+        "accuracy": 0.9
+      },
+      "eval_f1": {
+        "f1": 0.9008866348649631
+      },
+      "eval_loss": 0.4273129999637604,
+      "eval_precision": {
+        "precision": 0.9097813011730537
+      },
+      "eval_recall": {
+        "recall": 0.9019045670641415
+      },
+      "eval_runtime": 10.0655,
+      "eval_samples_per_second": 50.668,
+      "eval_steps_per_second": 25.334,
+      "step": 4335
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 2.403776325344953e-05,
+      "loss": 0.0162,
+      "step": 4340
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 2.3965141612200436e-05,
+      "loss": 0.0826,
+      "step": 4350
+    },
+    {
+      "epoch": 17.1,
+      "learning_rate": 2.3892519970951344e-05,
+      "loss": 0.1239,
+      "step": 4360
+    },
+    {
+      "epoch": 17.14,
+      "learning_rate": 2.3819898329702252e-05,
+      "loss": 0.1231,
+      "step": 4370
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 2.374727668845316e-05,
+      "loss": 0.0332,
+      "step": 4380
+    },
+    {
+      "epoch": 17.22,
+      "learning_rate": 2.3674655047204068e-05,
+      "loss": 0.1999,
+      "step": 4390
+    },
+    {
+      "epoch": 17.25,
+      "learning_rate": 2.3602033405954976e-05,
+      "loss": 0.0078,
+      "step": 4400
+    },
+    {
+      "epoch": 17.29,
+      "learning_rate": 2.3529411764705884e-05,
+      "loss": 0.0337,
+      "step": 4410
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 2.345679012345679e-05,
+      "loss": 0.0086,
+      "step": 4420
+    },
+    {
+      "epoch": 17.37,
+      "learning_rate": 2.33841684822077e-05,
+      "loss": 0.1673,
+      "step": 4430
+    },
+    {
+      "epoch": 17.41,
+      "learning_rate": 2.3311546840958608e-05,
+      "loss": 0.0645,
+      "step": 4440
+    },
+    {
+      "epoch": 17.45,
+      "learning_rate": 2.3238925199709515e-05,
+      "loss": 0.0929,
+      "step": 4450
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 2.3166303558460423e-05,
+      "loss": 0.3215,
+      "step": 4460
+    },
+    {
+      "epoch": 17.53,
+      "learning_rate": 2.3093681917211328e-05,
+      "loss": 0.2456,
+      "step": 4470
+    },
+    {
+      "epoch": 17.57,
+      "learning_rate": 2.302106027596224e-05,
+      "loss": 0.227,
+      "step": 4480
+    },
+    {
+      "epoch": 17.61,
+      "learning_rate": 2.2948438634713147e-05,
+      "loss": 0.0895,
+      "step": 4490
+    },
+    {
+      "epoch": 17.65,
+      "learning_rate": 2.2875816993464052e-05,
+      "loss": 0.0703,
+      "step": 4500
+    },
+    {
+      "epoch": 17.69,
+      "learning_rate": 2.2803195352214963e-05,
+      "loss": 0.1512,
+      "step": 4510
+    },
+    {
+      "epoch": 17.73,
+      "learning_rate": 2.2730573710965868e-05,
+      "loss": 0.007,
+      "step": 4520
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 2.2657952069716776e-05,
+      "loss": 0.1733,
+      "step": 4530
+    },
+    {
+      "epoch": 17.8,
+      "learning_rate": 2.2585330428467687e-05,
+      "loss": 0.0054,
+      "step": 4540
+    },
+    {
+      "epoch": 17.84,
+      "learning_rate": 2.251270878721859e-05,
+      "loss": 0.1822,
+      "step": 4550
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 2.24400871459695e-05,
+      "loss": 0.181,
+      "step": 4560
+    },
+    {
+      "epoch": 17.92,
+      "learning_rate": 2.2367465504720407e-05,
+      "loss": 0.0488,
+      "step": 4570
+    },
+    {
+      "epoch": 17.96,
+      "learning_rate": 2.2294843863471315e-05,
+      "loss": 0.1572,
+      "step": 4580
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.0727,
+      "step": 4590
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": {
+        "accuracy": 0.907843137254902
+      },
+      "eval_f1": {
+        "f1": 0.908247672636611
+      },
+      "eval_loss": 0.42159906029701233,
+      "eval_precision": {
+        "precision": 0.9152803626027508
+      },
+      "eval_recall": {
+        "recall": 0.908746882550074
+      },
+      "eval_runtime": 10.0325,
+      "eval_samples_per_second": 50.835,
+      "eval_steps_per_second": 25.418,
+      "step": 4590
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 7650,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2.845582368754729e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/tmp-checkpoint-4590/training_args.bin b/tmp-checkpoint-4590/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..dad10bac6a7c975c8619b9fbfae967cc5e44b76d
--- /dev/null
+++ b/tmp-checkpoint-4590/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418b85f35316fefe4bc777808f9dff13bb59b607bf93a8504dc8b5709313507f
+size 4728