End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7608695652173914
+      value: 0.8478260869565217
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.2530
-- Accuracy: 0.7609
+- Loss: 0.7770
+- Accuracy: 0.8478
 
 ## Model description
 

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 39.61,
-    "eval_accuracy": 0.8043478260869565,
-    "eval_loss": 2.584726139752747e+28,
-    "eval_runtime": 0.9982,
-    "eval_samples_per_second": 46.082,
-    "eval_steps_per_second": 6.011,
-    "train_loss": 6.009952755433709e+28,
-    "train_runtime": 674.9098,
-    "train_samples_per_second": 48.54,
-    "train_steps_per_second": 3.023
+    "eval_accuracy": 0.8478260869565217,
+    "eval_loss": 0.7770432233810425,
+    "eval_runtime": 1.0732,
+    "eval_samples_per_second": 42.861,
+    "eval_steps_per_second": 5.591,
+    "train_loss": 0.18828091986349546,
+    "train_runtime": 647.6496,
+    "train_samples_per_second": 50.583,
+    "train_steps_per_second": 3.15
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 39.61,
-    "eval_accuracy": 0.8043478260869565,
-    "eval_loss": 2.584726139752747e+28,
-    "eval_runtime": 0.9982,
-    "eval_samples_per_second": 46.082,
-    "eval_steps_per_second": 6.011
+    "eval_accuracy": 0.8478260869565217,
+    "eval_loss": 0.7770432233810425,
+    "eval_runtime": 1.0732,
+    "eval_samples_per_second": 42.861,
+    "eval_steps_per_second": 5.591
 }

runs/Jun23_16-08-59_DESKTOP-SKBE9FB/events.out.tfevents.1719181356.DESKTOP-SKBE9FB.16180.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edf082c17080765a531d7a3d13bcc101e1ec85b83d9bcc549e7a0db55add628f
+size 411

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 39.61,
-    "train_loss": 6.009952755433709e+28,
-    "train_runtime": 674.9098,
-    "train_samples_per_second": 48.54,
-    "train_steps_per_second": 3.023
+    "train_loss": 0.18828091986349546,
+    "train_runtime": 647.6496,
+    "train_samples_per_second": 50.583,
+    "train_steps_per_second": 3.15
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.8043478260869565,
-  "best_model_checkpoint": "vit-base-patch16-224-ve-U13b-80RX1\\checkpoint-360",
+  "best_metric": 0.8478260869565217,
+  "best_model_checkpoint": "vit-base-patch16-224-ve-U13b-80RX1\\checkpoint-824",
   "epoch": 39.61165048543689,
   "eval_steps": 500,
   "global_step": 2040,
@@ -11,37 +11,37 @@
     {
       "epoch": 0.19,
       "learning_rate": 5.392156862745098e-06,
-      "loss": 6.687978618025639e+28,
+      "loss": 1.3862,
       "step": 10
     },
     {
       "epoch": 0.39,
       "learning_rate": 1.0784313725490196e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 1.3853,
       "step": 20
     },
     {
       "epoch": 0.58,
       "learning_rate": 1.6176470588235296e-05,
-      "loss": 6.539356300076766e+28,
+      "loss": 1.3787,
       "step": 30
     },
     {
       "epoch": 0.78,
       "learning_rate": 2.156862745098039e-05,
-      "loss": 5.053138787427811e+28,
+      "loss": 1.3618,
       "step": 40
     },
     {
       "epoch": 0.97,
       "learning_rate": 2.696078431372549e-05,
-      "loss": 6.242113175336294e+28,
+      "loss": 1.3157,
       "step": 50
     },
     {
       "epoch": 0.99,
-      "eval_accuracy": 0.3695652173913043,
-      "eval_loss": 2.584726139752747e+28,
+      "eval_accuracy": 0.34782608695652173,
+      "eval_loss": 1.2967547178268433,
       "eval_runtime": 0.8432,
       "eval_samples_per_second": 54.555,
       "eval_steps_per_second": 7.116,
@@ -50,1556 +50,1556 @@
     {
       "epoch": 1.17,
       "learning_rate": 3.235294117647059e-05,
-      "loss": 5.300842021027069e+28,
+      "loss": 1.268,
       "step": 60
     },
     {
       "epoch": 1.36,
       "learning_rate": 3.774509803921569e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 1.1883,
       "step": 70
     },
     {
       "epoch": 1.55,
       "learning_rate": 4.313725490196078e-05,
-      "loss": 4.90451835842553e+28,
+      "loss": 1.1403,
       "step": 80
     },
     {
       "epoch": 1.75,
       "learning_rate": 4.8529411764705885e-05,
-      "loss": 7.133844060714986e+28,
+      "loss": 1.0174,
       "step": 90
     },
     {
       "epoch": 1.94,
       "learning_rate": 5.392156862745098e-05,
-      "loss": 6.836600935974513e+28,
+      "loss": 1.0334,
       "step": 100
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.5434782608695652,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7792,
-      "eval_samples_per_second": 59.035,
-      "eval_steps_per_second": 7.7,
+      "eval_accuracy": 0.5217391304347826,
+      "eval_loss": 1.0059559345245361,
+      "eval_runtime": 0.8107,
+      "eval_samples_per_second": 56.741,
+      "eval_steps_per_second": 7.401,
       "step": 103
     },
     {
       "epoch": 2.14,
       "learning_rate": 5.477296181630547e-05,
-      "loss": 5.152220836446151e+28,
+      "loss": 0.9461,
       "step": 110
     },
     {
       "epoch": 2.33,
       "learning_rate": 5.448916408668731e-05,
-      "loss": 6.985221742766112e+28,
+      "loss": 0.9687,
       "step": 120
     },
     {
       "epoch": 2.52,
       "learning_rate": 5.4205366357069146e-05,
-      "loss": 6.3907354932851674e+28,
+      "loss": 0.8107,
       "step": 130
     },
     {
       "epoch": 2.72,
       "learning_rate": 5.392156862745098e-05,
-      "loss": 5.350383045536239e+28,
+      "loss": 0.6984,
       "step": 140
     },
     {
       "epoch": 2.91,
       "learning_rate": 5.363777089783282e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.691,
       "step": 150
     },
     {
       "epoch": 2.99,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8532,
-      "eval_samples_per_second": 53.915,
-      "eval_steps_per_second": 7.032,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.75058913230896,
+      "eval_runtime": 0.8087,
+      "eval_samples_per_second": 56.882,
+      "eval_steps_per_second": 7.419,
       "step": 154
     },
     {
       "epoch": 3.11,
       "learning_rate": 5.3353973168214655e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.6486,
       "step": 160
     },
     {
       "epoch": 3.3,
       "learning_rate": 5.3070175438596496e-05,
-      "loss": 6.985223253923386e+28,
+      "loss": 0.5011,
       "step": 170
     },
     {
       "epoch": 3.5,
       "learning_rate": 5.278637770897833e-05,
-      "loss": 6.242113175336294e+28,
+      "loss": 0.5658,
       "step": 180
     },
     {
       "epoch": 3.69,
       "learning_rate": 5.2502579979360165e-05,
-      "loss": 5.499004230117157e+28,
+      "loss": 0.6253,
       "step": 190
     },
     {
       "epoch": 3.88,
       "learning_rate": 5.2218782249742006e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.5005,
       "step": 200
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8142,
-      "eval_samples_per_second": 56.498,
-      "eval_steps_per_second": 7.369,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.643278181552887,
+      "eval_runtime": 0.8062,
+      "eval_samples_per_second": 57.059,
+      "eval_steps_per_second": 7.443,
       "step": 206
     },
     {
       "epoch": 4.08,
       "learning_rate": 5.193498452012384e-05,
-      "loss": 5.746707463716415e+28,
+      "loss": 0.4162,
       "step": 210
     },
     {
       "epoch": 4.27,
       "learning_rate": 5.165118679050568e-05,
-      "loss": 6.3907354932851674e+28,
+      "loss": 0.3671,
       "step": 220
     },
     {
       "epoch": 4.47,
       "learning_rate": 5.1367389060887515e-05,
-      "loss": 6.539355544498129e+28,
+      "loss": 0.4731,
       "step": 230
     },
     {
       "epoch": 4.66,
       "learning_rate": 5.108359133126935e-05,
-      "loss": 6.3907354932851674e+28,
+      "loss": 0.4894,
       "step": 240
     },
     {
       "epoch": 4.85,
       "learning_rate": 5.079979360165119e-05,
-      "loss": 5.796248110436267e+28,
+      "loss": 0.3478,
       "step": 250
     },
     {
       "epoch": 4.99,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8212,
-      "eval_samples_per_second": 56.017,
-      "eval_steps_per_second": 7.307,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.5674107074737549,
+      "eval_runtime": 0.8242,
+      "eval_samples_per_second": 55.813,
+      "eval_steps_per_second": 7.28,
       "step": 257
     },
     {
       "epoch": 5.05,
       "learning_rate": 5.0515995872033025e-05,
-      "loss": 4.161409035417075e+28,
+      "loss": 0.273,
       "step": 260
     },
     {
       "epoch": 5.24,
       "learning_rate": 5.023219814241486e-05,
-      "loss": 6.539357811234041e+28,
+      "loss": 0.2627,
       "step": 270
     },
     {
       "epoch": 5.44,
       "learning_rate": 4.99484004127967e-05,
-      "loss": 6.836600935974513e+28,
+      "loss": 0.2588,
       "step": 280
     },
     {
       "epoch": 5.63,
       "learning_rate": 4.9664602683178534e-05,
-      "loss": 6.539356300076766e+28,
+      "loss": 0.2494,
       "step": 290
     },
     {
       "epoch": 5.83,
       "learning_rate": 4.9380804953560375e-05,
-      "loss": 5.647626925855348e+28,
+      "loss": 0.3339,
       "step": 300
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7827,
-      "eval_samples_per_second": 58.772,
-      "eval_steps_per_second": 7.666,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.6622583270072937,
+      "eval_runtime": 0.7892,
+      "eval_samples_per_second": 58.287,
+      "eval_steps_per_second": 7.603,
       "step": 309
     },
     {
       "epoch": 6.02,
       "learning_rate": 4.909700722394221e-05,
-      "loss": 4.5577334535972505e+28,
+      "loss": 0.2956,
       "step": 310
     },
     {
       "epoch": 6.21,
       "learning_rate": 4.8813209494324044e-05,
-      "loss": 5.647626170276711e+28,
+      "loss": 0.3412,
       "step": 320
     },
     {
       "epoch": 6.41,
       "learning_rate": 4.8529411764705885e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.3145,
       "step": 330
     },
     {
       "epoch": 6.6,
       "learning_rate": 4.824561403508772e-05,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.2888,
       "step": 340
     },
     {
       "epoch": 6.8,
       "learning_rate": 4.796181630546956e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.2232,
       "step": 350
     },
     {
       "epoch": 6.99,
       "learning_rate": 4.7678018575851394e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.2533,
       "step": 360
     },
     {
       "epoch": 6.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7872,
-      "eval_samples_per_second": 58.435,
-      "eval_steps_per_second": 7.622,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 0.6905426979064941,
+      "eval_runtime": 0.8007,
+      "eval_samples_per_second": 57.451,
+      "eval_steps_per_second": 7.494,
       "step": 360
     },
     {
       "epoch": 7.18,
       "learning_rate": 4.739422084623323e-05,
-      "loss": 6.48981678672487e+28,
+      "loss": 0.2497,
       "step": 370
     },
     {
       "epoch": 7.38,
       "learning_rate": 4.711042311661507e-05,
-      "loss": 5.053139920795767e+28,
+      "loss": 0.3348,
       "step": 380
     },
     {
       "epoch": 7.57,
       "learning_rate": 4.6826625386996904e-05,
-      "loss": 6.093492368544695e+28,
+      "loss": 0.2502,
       "step": 390
     },
     {
       "epoch": 7.77,
       "learning_rate": 4.6542827657378745e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.2309,
       "step": 400
     },
     {
       "epoch": 7.96,
       "learning_rate": 4.625902992776058e-05,
-      "loss": 6.985221742766112e+28,
+      "loss": 0.138,
       "step": 410
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8117,
-      "eval_samples_per_second": 56.672,
-      "eval_steps_per_second": 7.392,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.7250736355781555,
+      "eval_runtime": 0.7887,
+      "eval_samples_per_second": 58.325,
+      "eval_steps_per_second": 7.608,
       "step": 412
     },
     {
       "epoch": 8.16,
       "learning_rate": 4.597523219814241e-05,
-      "loss": 6.3907347377065295e+28,
+      "loss": 0.1209,
       "step": 420
     },
     {
       "epoch": 8.35,
       "learning_rate": 4.5691434468524254e-05,
-      "loss": 8.025574946093677e+28,
+      "loss": 0.0895,
       "step": 430
     },
     {
       "epoch": 8.54,
       "learning_rate": 4.540763673890609e-05,
-      "loss": 4.0127867174682015e+28,
+      "loss": 0.1707,
       "step": 440
     },
     {
       "epoch": 8.74,
       "learning_rate": 4.512383900928793e-05,
-      "loss": 5.499005363485113e+28,
+      "loss": 0.1935,
       "step": 450
     },
     {
       "epoch": 8.93,
       "learning_rate": 4.4840041279669764e-05,
-      "loss": 5.94487042838514e+28,
+      "loss": 0.1289,
       "step": 460
     },
     {
       "epoch": 8.99,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7912,
-      "eval_samples_per_second": 58.141,
-      "eval_steps_per_second": 7.584,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 0.7467479109764099,
+      "eval_runtime": 0.7842,
+      "eval_samples_per_second": 58.659,
+      "eval_steps_per_second": 7.651,
       "step": 463
     },
     {
       "epoch": 9.13,
       "learning_rate": 4.45562435500516e-05,
-      "loss": 5.300842776605706e+28,
+      "loss": 0.0955,
       "step": 470
     },
     {
       "epoch": 9.32,
       "learning_rate": 4.427244582043344e-05,
-      "loss": 5.647626925855348e+28,
+      "loss": 0.1515,
       "step": 480
     },
     {
       "epoch": 9.51,
       "learning_rate": 4.398864809081527e-05,
-      "loss": 6.539357811234041e+28,
+      "loss": 0.2114,
       "step": 490
     },
     {
       "epoch": 9.71,
       "learning_rate": 4.3704850361197114e-05,
-      "loss": 5.053139920795767e+28,
+      "loss": 0.0635,
       "step": 500
     },
     {
       "epoch": 9.9,
       "learning_rate": 4.342105263157895e-05,
-      "loss": 8.174196508463913e+28,
+      "loss": 0.152,
       "step": 510
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8382,
-      "eval_samples_per_second": 54.88,
-      "eval_steps_per_second": 7.158,
+      "eval_accuracy": 0.717391304347826,
+      "eval_loss": 0.9011275172233582,
+      "eval_runtime": 0.8162,
+      "eval_samples_per_second": 56.36,
+      "eval_steps_per_second": 7.351,
       "step": 515
     },
     {
       "epoch": 10.1,
       "learning_rate": 4.313725490196078e-05,
-      "loss": 5.449463961186624e+28,
+      "loss": 0.113,
       "step": 520
     },
     {
       "epoch": 10.29,
       "learning_rate": 4.2853457172342624e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.1291,
       "step": 530
     },
     {
       "epoch": 10.49,
       "learning_rate": 4.2569659442724465e-05,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.1063,
       "step": 540
     },
     {
       "epoch": 10.68,
       "learning_rate": 4.22858617131063e-05,
-      "loss": 6.3907347377065295e+28,
+      "loss": 0.1218,
       "step": 550
     },
     {
       "epoch": 10.87,
       "learning_rate": 4.200206398348813e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 0.2609,
       "step": 560
     },
     {
       "epoch": 10.99,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7817,
-      "eval_samples_per_second": 58.847,
-      "eval_steps_per_second": 7.676,
+      "eval_accuracy": 0.717391304347826,
+      "eval_loss": 1.0149883031845093,
+      "eval_runtime": 0.8082,
+      "eval_samples_per_second": 56.918,
+      "eval_steps_per_second": 7.424,
       "step": 566
     },
     {
       "epoch": 11.07,
       "learning_rate": 4.171826625386997e-05,
-      "loss": 5.5980859013461785e+28,
+      "loss": 0.2021,
       "step": 570
     },
     {
       "epoch": 11.26,
       "learning_rate": 4.143446852425181e-05,
-      "loss": 6.242113553125612e+28,
+      "loss": 0.1712,
       "step": 580
     },
     {
       "epoch": 11.46,
       "learning_rate": 4.115067079463365e-05,
-      "loss": 7.4310871854554575e+28,
+      "loss": 0.0442,
       "step": 590
     },
     {
       "epoch": 11.65,
       "learning_rate": 4.0866873065015484e-05,
-      "loss": 5.350383045536239e+28,
+      "loss": 0.1487,
       "step": 600
     },
     {
       "epoch": 11.84,
       "learning_rate": 4.058307533539732e-05,
-      "loss": 5.796247354857629e+28,
+      "loss": 0.2202,
       "step": 610
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7862,
-      "eval_samples_per_second": 58.51,
-      "eval_steps_per_second": 7.632,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.9713221788406372,
+      "eval_runtime": 0.7902,
+      "eval_samples_per_second": 58.215,
+      "eval_steps_per_second": 7.593,
       "step": 618
     },
     {
       "epoch": 12.04,
       "learning_rate": 4.029927760577915e-05,
-      "loss": 4.90451835842553e+28,
+      "loss": 0.1409,
       "step": 620
     },
     {
       "epoch": 12.23,
       "learning_rate": 4.001547987616099e-05,
-      "loss": 5.053139920795767e+28,
+      "loss": 0.1339,
       "step": 630
     },
     {
       "epoch": 12.43,
       "learning_rate": 3.9731682146542834e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.1061,
       "step": 640
     },
     {
       "epoch": 12.62,
       "learning_rate": 3.944788441692467e-05,
-      "loss": 7.282465623085222e+28,
+      "loss": 0.0665,
       "step": 650
     },
     {
       "epoch": 12.82,
       "learning_rate": 3.91640866873065e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 0.1083,
       "step": 660
     },
     {
       "epoch": 12.99,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7922,
-      "eval_samples_per_second": 58.067,
-      "eval_steps_per_second": 7.574,
+      "eval_accuracy": 0.6739130434782609,
+      "eval_loss": 1.1106468439102173,
+      "eval_runtime": 0.8002,
+      "eval_samples_per_second": 57.487,
+      "eval_steps_per_second": 7.498,
       "step": 669
     },
     {
       "epoch": 13.01,
       "learning_rate": 3.888028895768834e-05,
-      "loss": 5.647626170276711e+28,
+      "loss": 0.0472,
       "step": 670
     },
     {
       "epoch": 13.2,
       "learning_rate": 3.859649122807018e-05,
-      "loss": 6.687979373604276e+28,
+      "loss": 0.1036,
       "step": 680
     },
     {
       "epoch": 13.4,
       "learning_rate": 3.831269349845202e-05,
-      "loss": 6.985221742766112e+28,
+      "loss": 0.1185,
       "step": 690
     },
     {
       "epoch": 13.59,
       "learning_rate": 3.802889576883385e-05,
-      "loss": 4.458652537946866e+28,
+      "loss": 0.0721,
       "step": 700
     },
     {
       "epoch": 13.79,
       "learning_rate": 3.774509803921569e-05,
-      "loss": 6.093492368544695e+28,
+      "loss": 0.1805,
       "step": 710
     },
     {
       "epoch": 13.98,
       "learning_rate": 3.746130030959752e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.07,
       "step": 720
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7842,
-      "eval_samples_per_second": 58.659,
-      "eval_steps_per_second": 7.651,
+      "eval_accuracy": 0.717391304347826,
+      "eval_loss": 1.121076226234436,
+      "eval_runtime": 0.8312,
+      "eval_samples_per_second": 55.342,
+      "eval_steps_per_second": 7.219,
       "step": 721
     },
     {
       "epoch": 14.17,
       "learning_rate": 3.7177502579979356e-05,
-      "loss": 5.449464338975942e+28,
+      "loss": 0.0911,
       "step": 730
     },
     {
       "epoch": 14.37,
       "learning_rate": 3.6893704850361204e-05,
-      "loss": 6.242113175336294e+28,
+      "loss": 0.1052,
       "step": 740
     },
     {
       "epoch": 14.56,
       "learning_rate": 3.660990712074304e-05,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.1289,
       "step": 750
     },
     {
       "epoch": 14.76,
       "learning_rate": 3.632610939112487e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.118,
       "step": 760
     },
     {
       "epoch": 14.95,
       "learning_rate": 3.6042311661506706e-05,
-      "loss": 5.796247732646948e+28,
+      "loss": 0.0791,
       "step": 770
     },
     {
       "epoch": 14.99,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8192,
-      "eval_samples_per_second": 56.153,
-      "eval_steps_per_second": 7.324,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.1829625368118286,
+      "eval_runtime": 0.8107,
+      "eval_samples_per_second": 56.742,
+      "eval_steps_per_second": 7.401,
       "step": 772
     },
     {
       "epoch": 15.15,
       "learning_rate": 3.575851393188854e-05,
-      "loss": 7.183384329645518e+28,
+      "loss": 0.0461,
       "step": 780
     },
     {
       "epoch": 15.34,
       "learning_rate": 3.547471620227039e-05,
-      "loss": 5.944870806174458e+28,
+      "loss": 0.2331,
       "step": 790
     },
     {
       "epoch": 15.53,
       "learning_rate": 3.519091847265222e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.0504,
       "step": 800
     },
     {
       "epoch": 15.73,
       "learning_rate": 3.490712074303406e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.0089,
       "step": 810
     },
     {
       "epoch": 15.92,
       "learning_rate": 3.462332301341589e-05,
-      "loss": 5.0531391652171296e+28,
+      "loss": 0.0427,
       "step": 820
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7942,
-      "eval_samples_per_second": 57.921,
-      "eval_steps_per_second": 7.555,
+      "eval_accuracy": 0.8478260869565217,
+      "eval_loss": 0.7770432233810425,
+      "eval_runtime": 0.7982,
+      "eval_samples_per_second": 57.631,
+      "eval_steps_per_second": 7.517,
       "step": 824
     },
     {
       "epoch": 16.12,
       "learning_rate": 3.4339525283797725e-05,
-      "loss": 7.4310871854554575e+28,
+      "loss": 0.1128,
       "step": 830
     },
     {
       "epoch": 16.31,
       "learning_rate": 3.405572755417957e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.1035,
       "step": 840
     },
     {
       "epoch": 16.5,
       "learning_rate": 3.377192982456141e-05,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.0311,
       "step": 850
     },
     {
       "epoch": 16.7,
       "learning_rate": 3.348813209494324e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0639,
       "step": 860
     },
     {
       "epoch": 16.89,
       "learning_rate": 3.3204334365325076e-05,
-      "loss": 5.796248110436267e+28,
+      "loss": 0.1219,
       "step": 870
     },
     {
       "epoch": 16.99,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7732,
-      "eval_samples_per_second": 59.495,
-      "eval_steps_per_second": 7.76,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 1.096178650856018,
+      "eval_runtime": 0.7962,
+      "eval_samples_per_second": 57.776,
+      "eval_steps_per_second": 7.536,
       "step": 875
     },
     {
       "epoch": 17.09,
       "learning_rate": 3.292053663570691e-05,
-      "loss": 5.300842776605706e+28,
+      "loss": 0.0854,
       "step": 880
     },
     {
       "epoch": 17.28,
       "learning_rate": 3.263673890608876e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.1007,
       "step": 890
     },
     {
       "epoch": 17.48,
       "learning_rate": 3.235294117647059e-05,
-      "loss": 7.4310879410340955e+28,
+      "loss": 0.0978,
       "step": 900
     },
     {
       "epoch": 17.67,
       "learning_rate": 3.2069143446852426e-05,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.125,
       "step": 910
     },
     {
       "epoch": 17.86,
       "learning_rate": 3.178534571723426e-05,
-      "loss": 4.755896040476657e+28,
+      "loss": 0.0739,
       "step": 920
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8047,
-      "eval_samples_per_second": 57.165,
-      "eval_steps_per_second": 7.456,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 0.9447097182273865,
+      "eval_runtime": 0.8317,
+      "eval_samples_per_second": 55.309,
+      "eval_steps_per_second": 7.214,
       "step": 927
     },
     {
       "epoch": 18.06,
       "learning_rate": 3.1501547987616095e-05,
-      "loss": 6.935681473835579e+28,
+      "loss": 0.0827,
       "step": 930
     },
     {
       "epoch": 18.25,
       "learning_rate": 3.121775025799794e-05,
-      "loss": 5.94487042838514e+28,
+      "loss": 0.0488,
       "step": 940
     },
     {
       "epoch": 18.45,
       "learning_rate": 3.093395252837978e-05,
-      "loss": 5.796247732646948e+28,
+      "loss": 0.1147,
       "step": 950
     },
     {
       "epoch": 18.64,
       "learning_rate": 3.065015479876161e-05,
-      "loss": 5.944869672806502e+28,
+      "loss": 0.0895,
       "step": 960
     },
     {
       "epoch": 18.83,
       "learning_rate": 3.0366357069143445e-05,
-      "loss": 6.836600180395876e+28,
+      "loss": 0.1989,
       "step": 970
     },
     {
       "epoch": 18.99,
-      "eval_accuracy": 0.782608695652174,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8292,
-      "eval_samples_per_second": 55.476,
-      "eval_steps_per_second": 7.236,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 1.1543340682983398,
+      "eval_runtime": 0.8057,
+      "eval_samples_per_second": 57.094,
+      "eval_steps_per_second": 7.447,
       "step": 978
     },
     {
       "epoch": 19.03,
       "learning_rate": 3.0082559339525283e-05,
-      "loss": 5.746708219295052e+28,
+      "loss": 0.0619,
       "step": 980
     },
     {
       "epoch": 19.22,
       "learning_rate": 2.9798761609907124e-05,
-      "loss": 5.3503834233255575e+28,
+      "loss": 0.0551,
       "step": 990
     },
     {
       "epoch": 19.42,
       "learning_rate": 2.9514963880288958e-05,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.0815,
       "step": 1000
     },
     {
       "epoch": 19.61,
       "learning_rate": 2.9231166150670796e-05,
-      "loss": 6.687979373604276e+28,
+      "loss": 0.1433,
       "step": 1010
     },
     {
       "epoch": 19.81,
       "learning_rate": 2.894736842105263e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.0481,
       "step": 1020
     },
     {
       "epoch": 20.0,
       "learning_rate": 2.8663570691434468e-05,
-      "loss": 6.043951721824843e+28,
+      "loss": 0.1097,
       "step": 1030
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7862,
-      "eval_samples_per_second": 58.511,
-      "eval_steps_per_second": 7.632,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.1795116662979126,
+      "eval_runtime": 0.8007,
+      "eval_samples_per_second": 57.45,
+      "eval_steps_per_second": 7.493,
       "step": 1030
     },
     {
       "epoch": 20.19,
       "learning_rate": 2.837977296181631e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 0.1128,
       "step": 1040
     },
     {
       "epoch": 20.39,
       "learning_rate": 2.8095975232198143e-05,
-      "loss": 6.093491990755376e+28,
+      "loss": 0.1016,
       "step": 1050
     },
     {
       "epoch": 20.58,
       "learning_rate": 2.781217750257998e-05,
-      "loss": 5.3503834233255575e+28,
+      "loss": 0.1064,
       "step": 1060
     },
     {
       "epoch": 20.78,
       "learning_rate": 2.7528379772961815e-05,
-      "loss": 5.647626925855348e+28,
+      "loss": 0.1248,
       "step": 1070
     },
     {
       "epoch": 20.97,
       "learning_rate": 2.7244582043343656e-05,
-      "loss": 7.579709503404331e+28,
+      "loss": 0.1204,
       "step": 1080
     },
     {
       "epoch": 20.99,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8627,
-      "eval_samples_per_second": 53.321,
-      "eval_steps_per_second": 6.955,
+      "eval_accuracy": 0.6739130434782609,
+      "eval_loss": 1.267902135848999,
+      "eval_runtime": 0.7942,
+      "eval_samples_per_second": 57.922,
+      "eval_steps_per_second": 7.555,
       "step": 1081
     },
     {
       "epoch": 21.17,
       "learning_rate": 2.696078431372549e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0909,
       "step": 1090
     },
     {
       "epoch": 21.36,
       "learning_rate": 2.6676986584107328e-05,
-      "loss": 4.904517602846893e+28,
+      "loss": 0.0697,
       "step": 1100
     },
     {
       "epoch": 21.55,
       "learning_rate": 2.6393188854489165e-05,
-      "loss": 6.242113175336294e+28,
+      "loss": 0.0691,
       "step": 1110
     },
     {
       "epoch": 21.75,
       "learning_rate": 2.6109391124871003e-05,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.0676,
       "step": 1120
     },
     {
       "epoch": 21.94,
       "learning_rate": 2.582559339525284e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0514,
       "step": 1130
     },
     {
       "epoch": 22.0,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7952,
-      "eval_samples_per_second": 57.847,
-      "eval_steps_per_second": 7.545,
+      "eval_accuracy": 0.717391304347826,
+      "eval_loss": 1.0645781755447388,
+      "eval_runtime": 0.7947,
+      "eval_samples_per_second": 57.884,
+      "eval_steps_per_second": 7.55,
       "step": 1133
     },
     {
       "epoch": 22.14,
       "learning_rate": 2.5541795665634675e-05,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.0133,
       "step": 1140
     },
     {
       "epoch": 22.33,
       "learning_rate": 2.5257997936016512e-05,
-      "loss": 5.796248866014904e+28,
+      "loss": 0.051,
       "step": 1150
     },
     {
       "epoch": 22.52,
       "learning_rate": 2.497420020639835e-05,
-      "loss": 5.0531391652171296e+28,
+      "loss": 0.1221,
       "step": 1160
     },
     {
       "epoch": 22.72,
       "learning_rate": 2.4690402476780188e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 0.0973,
       "step": 1170
     },
     {
       "epoch": 22.91,
       "learning_rate": 2.4406604747162022e-05,
-      "loss": 6.3907354932851674e+28,
+      "loss": 0.0612,
       "step": 1180
     },
     {
       "epoch": 22.99,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7682,
-      "eval_samples_per_second": 59.882,
-      "eval_steps_per_second": 7.811,
+      "eval_accuracy": 0.6956521739130435,
+      "eval_loss": 1.1412676572799683,
+      "eval_runtime": 0.8302,
+      "eval_samples_per_second": 55.409,
+      "eval_steps_per_second": 7.227,
       "step": 1184
     },
     {
       "epoch": 23.11,
       "learning_rate": 2.412280701754386e-05,
-      "loss": 5.449464338975942e+28,
+      "loss": 0.0519,
       "step": 1190
     },
     {
       "epoch": 23.3,
       "learning_rate": 2.3839009287925697e-05,
-      "loss": 5.944869672806502e+28,
+      "loss": 0.0405,
       "step": 1200
     },
     {
       "epoch": 23.5,
       "learning_rate": 2.3555211558307535e-05,
-      "loss": 5.796247732646948e+28,
+      "loss": 0.0487,
       "step": 1210
     },
     {
       "epoch": 23.69,
       "learning_rate": 2.3271413828689372e-05,
-      "loss": 5.350383045536239e+28,
+      "loss": 0.0784,
       "step": 1220
     },
     {
       "epoch": 23.88,
       "learning_rate": 2.2987616099071207e-05,
-      "loss": 7.876953383723441e+28,
+      "loss": 0.0207,
       "step": 1230
     },
     {
       "epoch": 24.0,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7822,
-      "eval_samples_per_second": 58.81,
-      "eval_steps_per_second": 7.671,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 0.8928263187408447,
+      "eval_runtime": 0.8367,
+      "eval_samples_per_second": 54.978,
+      "eval_steps_per_second": 7.171,
       "step": 1236
     },
     {
       "epoch": 24.08,
       "learning_rate": 2.2703818369453044e-05,
-      "loss": 6.6384375935164695e+28,
+      "loss": 0.0647,
       "step": 1240
     },
     {
       "epoch": 24.27,
       "learning_rate": 2.2420020639834882e-05,
-      "loss": 4.755895662687339e+28,
+      "loss": 0.1242,
       "step": 1250
     },
     {
       "epoch": 24.47,
       "learning_rate": 2.213622291021672e-05,
-      "loss": 5.499004985695794e+28,
+      "loss": 0.0921,
       "step": 1260
     },
     {
       "epoch": 24.66,
       "learning_rate": 2.1852425180598557e-05,
-      "loss": 5.944869295017184e+28,
+      "loss": 0.057,
       "step": 1270
     },
     {
       "epoch": 24.85,
       "learning_rate": 2.156862745098039e-05,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.1063,
       "step": 1280
     },
     {
       "epoch": 24.99,
-      "eval_accuracy": 0.8043478260869565,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7867,
-      "eval_samples_per_second": 58.473,
-      "eval_steps_per_second": 7.627,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.118552327156067,
+      "eval_runtime": 0.7882,
+      "eval_samples_per_second": 58.362,
+      "eval_steps_per_second": 7.612,
       "step": 1287
     },
     {
       "epoch": 25.05,
       "learning_rate": 2.1284829721362232e-05,
-      "loss": 6.8861404493264086e+28,
+      "loss": 0.0474,
       "step": 1290
     },
     {
       "epoch": 25.24,
       "learning_rate": 2.1001031991744067e-05,
-      "loss": 6.3907354932851674e+28,
+      "loss": 0.1243,
       "step": 1300
     },
     {
       "epoch": 25.44,
       "learning_rate": 2.0717234262125904e-05,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0265,
       "step": 1310
     },
     {
       "epoch": 25.63,
       "learning_rate": 2.0433436532507742e-05,
-      "loss": 6.985221742766112e+28,
+      "loss": 0.1075,
       "step": 1320
     },
     {
       "epoch": 25.83,
       "learning_rate": 2.0149638802889576e-05,
-      "loss": 5.647626925855348e+28,
+      "loss": 0.1076,
       "step": 1330
     },
     {
       "epoch": 26.0,
       "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7822,
-      "eval_samples_per_second": 58.81,
-      "eval_steps_per_second": 7.671,
+      "eval_loss": 1.1741024255752563,
+      "eval_runtime": 0.8087,
+      "eval_samples_per_second": 56.882,
+      "eval_steps_per_second": 7.419,
       "step": 1339
     },
     {
       "epoch": 26.02,
       "learning_rate": 1.9865841073271417e-05,
-      "loss": 5.15222121423547e+28,
+      "loss": 0.0272,
       "step": 1340
     },
     {
       "epoch": 26.21,
       "learning_rate": 1.958204334365325e-05,
-      "loss": 4.904517602846893e+28,
+      "loss": 0.0688,
       "step": 1350
     },
     {
       "epoch": 26.41,
       "learning_rate": 1.929824561403509e-05,
-      "loss": 7.133844060714986e+28,
+      "loss": 0.0913,
       "step": 1360
     },
     {
       "epoch": 26.6,
       "learning_rate": 1.9014447884416927e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0374,
       "step": 1370
     },
     {
       "epoch": 26.8,
       "learning_rate": 1.873065015479876e-05,
-      "loss": 5.796248110436267e+28,
+      "loss": 0.0891,
       "step": 1380
     },
     {
       "epoch": 26.99,
       "learning_rate": 1.8446852425180602e-05,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.0714,
       "step": 1390
     },
     {
       "epoch": 26.99,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8347,
-      "eval_samples_per_second": 55.11,
-      "eval_steps_per_second": 7.188,
+      "eval_accuracy": 0.8043478260869565,
+      "eval_loss": 1.0977226495742798,
+      "eval_runtime": 0.7997,
+      "eval_samples_per_second": 57.522,
+      "eval_steps_per_second": 7.503,
       "step": 1390
     },
     {
       "epoch": 27.18,
       "learning_rate": 1.8163054695562436e-05,
-      "loss": 6.78706066704398e+28,
+      "loss": 0.0417,
       "step": 1400
     },
     {
       "epoch": 27.38,
       "learning_rate": 1.787925696594427e-05,
-      "loss": 5.350383045536239e+28,
+      "loss": 0.1741,
       "step": 1410
     },
     {
       "epoch": 27.57,
       "learning_rate": 1.759545923632611e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0141,
       "step": 1420
     },
     {
       "epoch": 27.77,
       "learning_rate": 1.7311661506707946e-05,
-      "loss": 5.647626170276711e+28,
+      "loss": 0.0031,
       "step": 1430
     },
     {
       "epoch": 27.96,
       "learning_rate": 1.7027863777089787e-05,
-      "loss": 6.093491990755376e+28,
+      "loss": 0.062,
       "step": 1440
     },
     {
       "epoch": 28.0,
       "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7802,
-      "eval_samples_per_second": 58.961,
-      "eval_steps_per_second": 7.691,
+      "eval_loss": 1.3964738845825195,
+      "eval_runtime": 0.8222,
+      "eval_samples_per_second": 55.949,
+      "eval_steps_per_second": 7.298,
       "step": 1442
     },
     {
       "epoch": 28.16,
       "learning_rate": 1.674406604747162e-05,
-      "loss": 7.084303036205815e+28,
+      "loss": 0.1189,
       "step": 1450
     },
     {
       "epoch": 28.35,
       "learning_rate": 1.6460268317853455e-05,
-      "loss": 6.985222498344749e+28,
+      "loss": 0.085,
       "step": 1460
     },
     {
       "epoch": 28.54,
       "learning_rate": 1.6176470588235296e-05,
-      "loss": 4.458652160157547e+28,
+      "loss": 0.0294,
       "step": 1470
     },
     {
       "epoch": 28.74,
       "learning_rate": 1.589267285861713e-05,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.0267,
       "step": 1480
     },
     {
       "epoch": 28.93,
       "learning_rate": 1.560887512899897e-05,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0617,
       "step": 1490
     },
     {
       "epoch": 28.99,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7762,
-      "eval_samples_per_second": 59.264,
-      "eval_steps_per_second": 7.73,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.184866189956665,
+      "eval_runtime": 0.8482,
+      "eval_samples_per_second": 54.233,
+      "eval_steps_per_second": 7.074,
       "step": 1493
     },
     {
       "epoch": 29.13,
       "learning_rate": 1.5325077399380806e-05,
-      "loss": 5.895329403875969e+28,
+      "loss": 0.0499,
       "step": 1500
     },
     {
       "epoch": 29.32,
       "learning_rate": 1.5041279669762642e-05,
-      "loss": 5.944869672806502e+28,
+      "loss": 0.0568,
       "step": 1510
     },
     {
       "epoch": 29.51,
       "learning_rate": 1.4757481940144479e-05,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.0511,
       "step": 1520
     },
     {
       "epoch": 29.71,
       "learning_rate": 1.4473684210526315e-05,
-      "loss": 5.3503838011148765e+28,
+      "loss": 0.0284,
       "step": 1530
     },
     {
       "epoch": 29.9,
       "learning_rate": 1.4189886480908154e-05,
-      "loss": 6.985223253923386e+28,
+      "loss": 0.0536,
       "step": 1540
     },
     {
       "epoch": 30.0,
       "eval_accuracy": 0.782608695652174,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7882,
-      "eval_samples_per_second": 58.363,
-      "eval_steps_per_second": 7.613,
+      "eval_loss": 1.086539387702942,
+      "eval_runtime": 0.7992,
+      "eval_samples_per_second": 57.559,
+      "eval_steps_per_second": 7.508,
       "step": 1545
     },
     {
       "epoch": 30.1,
       "learning_rate": 1.390608875128999e-05,
-      "loss": 5.449463961186624e+28,
+      "loss": 0.0039,
       "step": 1550
     },
     {
       "epoch": 30.29,
       "learning_rate": 1.3622291021671828e-05,
-      "loss": 4.755896040476657e+28,
+      "loss": 0.038,
       "step": 1560
     },
     {
       "epoch": 30.49,
       "learning_rate": 1.3338493292053664e-05,
-      "loss": 5.796248488225585e+28,
+      "loss": 0.1408,
       "step": 1570
     },
     {
       "epoch": 30.68,
       "learning_rate": 1.3054695562435501e-05,
-      "loss": 6.836600935974513e+28,
+      "loss": 0.012,
       "step": 1580
     },
     {
       "epoch": 30.87,
       "learning_rate": 1.2770897832817337e-05,
-      "loss": 7.133844060714986e+28,
+      "loss": 0.0707,
       "step": 1590
     },
     {
       "epoch": 30.99,
       "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8897,
-      "eval_samples_per_second": 51.701,
-      "eval_steps_per_second": 6.744,
+      "eval_loss": 1.2080574035644531,
+      "eval_runtime": 0.8012,
+      "eval_samples_per_second": 57.415,
+      "eval_steps_per_second": 7.489,
       "step": 1596
     },
     {
       "epoch": 31.07,
       "learning_rate": 1.2487100103199175e-05,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0148,
       "step": 1600
     },
     {
       "epoch": 31.26,
       "learning_rate": 1.2203302373581011e-05,
-      "loss": 6.3907347377065295e+28,
+      "loss": 0.0309,
       "step": 1610
     },
     {
       "epoch": 31.46,
       "learning_rate": 1.1919504643962849e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0383,
       "step": 1620
     },
     {
       "epoch": 31.65,
       "learning_rate": 1.1635706914344686e-05,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.0231,
       "step": 1630
     },
     {
       "epoch": 31.84,
       "learning_rate": 1.1351909184726522e-05,
-      "loss": 6.687979373604276e+28,
+      "loss": 0.0967,
       "step": 1640
     },
     {
       "epoch": 32.0,
-      "eval_accuracy": 0.7608695652173914,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8212,
-      "eval_samples_per_second": 56.017,
-      "eval_steps_per_second": 7.307,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 1.330019474029541,
+      "eval_runtime": 0.8382,
+      "eval_samples_per_second": 54.88,
+      "eval_steps_per_second": 7.158,
       "step": 1648
     },
     {
       "epoch": 32.04,
       "learning_rate": 1.106811145510836e-05,
-      "loss": 5.449463961186624e+28,
+      "loss": 0.0414,
       "step": 1650
     },
     {
       "epoch": 32.23,
       "learning_rate": 1.0784313725490196e-05,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0196,
       "step": 1660
     },
     {
       "epoch": 32.43,
       "learning_rate": 1.0500515995872033e-05,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0517,
       "step": 1670
     },
     {
       "epoch": 32.62,
       "learning_rate": 1.0216718266253871e-05,
-      "loss": 7.7283318213532045e+28,
+      "loss": 0.0126,
       "step": 1680
     },
     {
       "epoch": 32.82,
       "learning_rate": 9.932920536635709e-06,
-      "loss": 4.458652537946866e+28,
+      "loss": 0.0564,
       "step": 1690
     },
     {
       "epoch": 32.99,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8102,
-      "eval_samples_per_second": 56.775,
-      "eval_steps_per_second": 7.405,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 1.2239521741867065,
+      "eval_runtime": 0.8475,
+      "eval_samples_per_second": 54.275,
+      "eval_steps_per_second": 7.079,
       "step": 1699
     },
     {
       "epoch": 33.01,
       "learning_rate": 9.649122807017545e-06,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.1062,
       "step": 1700
     },
     {
       "epoch": 33.2,
       "learning_rate": 9.36532507739938e-06,
-      "loss": 5.944869295017184e+28,
+      "loss": 0.0345,
       "step": 1710
     },
     {
       "epoch": 33.4,
       "learning_rate": 9.081527347781218e-06,
-      "loss": 6.539356300076766e+28,
+      "loss": 0.0488,
       "step": 1720
     },
     {
       "epoch": 33.59,
       "learning_rate": 8.797729618163056e-06,
-      "loss": 6.093491990755376e+28,
+      "loss": 0.0097,
       "step": 1730
     },
     {
       "epoch": 33.79,
       "learning_rate": 8.513931888544893e-06,
-      "loss": 5.796247732646948e+28,
+      "loss": 0.0395,
       "step": 1740
     },
     {
       "epoch": 33.98,
       "learning_rate": 8.230134158926728e-06,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0435,
       "step": 1750
     },
     {
       "epoch": 34.0,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8142,
-      "eval_samples_per_second": 56.495,
-      "eval_steps_per_second": 7.369,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.239098072052002,
+      "eval_runtime": 0.7892,
+      "eval_samples_per_second": 58.289,
+      "eval_steps_per_second": 7.603,
       "step": 1751
     },
     {
       "epoch": 34.17,
       "learning_rate": 7.946336429308565e-06,
-      "loss": 5.449464338975942e+28,
+      "loss": 0.0378,
       "step": 1760
     },
     {
       "epoch": 34.37,
       "learning_rate": 7.662538699690403e-06,
-      "loss": 6.985221742766112e+28,
+      "loss": 0.0318,
       "step": 1770
     },
     {
       "epoch": 34.56,
       "learning_rate": 7.3787409700722396e-06,
-      "loss": 4.904517225057574e+28,
+      "loss": 0.0482,
       "step": 1780
     },
     {
       "epoch": 34.76,
       "learning_rate": 7.094943240454077e-06,
-      "loss": 6.539356300076766e+28,
+      "loss": 0.01,
       "step": 1790
     },
     {
       "epoch": 34.95,
       "learning_rate": 6.811145510835914e-06,
-      "loss": 6.3907347377065295e+28,
+      "loss": 0.043,
       "step": 1800
     },
     {
       "epoch": 34.99,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7832,
-      "eval_samples_per_second": 58.735,
-      "eval_steps_per_second": 7.661,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.181269645690918,
+      "eval_runtime": 0.8162,
+      "eval_samples_per_second": 56.36,
+      "eval_steps_per_second": 7.351,
       "step": 1802
     },
     {
       "epoch": 35.15,
       "learning_rate": 6.527347781217751e-06,
-      "loss": 5.300842021027069e+28,
+      "loss": 0.0863,
       "step": 1810
     },
     {
       "epoch": 35.34,
       "learning_rate": 6.2435500515995875e-06,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0324,
       "step": 1820
     },
     {
       "epoch": 35.53,
       "learning_rate": 5.959752321981424e-06,
-      "loss": 6.539357055655403e+28,
+      "loss": 0.0119,
       "step": 1830
     },
     {
       "epoch": 35.73,
       "learning_rate": 5.675954592363261e-06,
-      "loss": 5.94487042838514e+28,
+      "loss": 0.0153,
       "step": 1840
     },
     {
       "epoch": 35.92,
       "learning_rate": 5.392156862745098e-06,
-      "loss": 7.579709503404331e+28,
+      "loss": 0.0218,
       "step": 1850
     },
     {
       "epoch": 36.0,
-      "eval_accuracy": 0.7391304347826086,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8372,
-      "eval_samples_per_second": 54.945,
-      "eval_steps_per_second": 7.167,
+      "eval_accuracy": 0.782608695652174,
+      "eval_loss": 1.2496311664581299,
+      "eval_runtime": 0.8557,
+      "eval_samples_per_second": 53.757,
+      "eval_steps_per_second": 7.012,
       "step": 1854
     },
     {
       "epoch": 36.12,
       "learning_rate": 5.1083591331269355e-06,
-      "loss": 4.90451835842553e+28,
+      "loss": 0.0449,
       "step": 1860
     },
     {
       "epoch": 36.31,
       "learning_rate": 4.824561403508772e-06,
-      "loss": 5.053139920795767e+28,
+      "loss": 0.0531,
       "step": 1870
     },
     {
       "epoch": 36.5,
       "learning_rate": 4.540763673890609e-06,
-      "loss": 7.72833031019593e+28,
+      "loss": 0.0336,
       "step": 1880
     },
     {
       "epoch": 36.7,
       "learning_rate": 4.256965944272447e-06,
-      "loss": 4.90451835842553e+28,
+      "loss": 0.1423,
       "step": 1890
     },
     {
       "epoch": 36.89,
       "learning_rate": 3.973168214654283e-06,
-      "loss": 6.242113553125612e+28,
+      "loss": 0.0043,
       "step": 1900
     },
     {
       "epoch": 36.99,
       "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8212,
-      "eval_samples_per_second": 56.017,
-      "eval_steps_per_second": 7.307,
+      "eval_loss": 1.2797400951385498,
+      "eval_runtime": 0.7897,
+      "eval_samples_per_second": 58.25,
+      "eval_steps_per_second": 7.598,
       "step": 1905
     },
     {
       "epoch": 37.09,
       "learning_rate": 3.6893704850361198e-06,
-      "loss": 7.232925354154688e+28,
+      "loss": 0.0707,
       "step": 1910
     },
     {
       "epoch": 37.28,
       "learning_rate": 3.405572755417957e-06,
-      "loss": 7.282466378663859e+28,
+      "loss": 0.0878,
       "step": 1920
     },
     {
       "epoch": 37.48,
       "learning_rate": 3.1217750257997938e-06,
-      "loss": 5.944870050595821e+28,
+      "loss": 0.0214,
       "step": 1930
     },
     {
       "epoch": 37.67,
       "learning_rate": 2.8379772961816305e-06,
-      "loss": 5.201761105376684e+28,
+      "loss": 0.0337,
       "step": 1940
     },
     {
       "epoch": 37.86,
       "learning_rate": 2.5541795665634677e-06,
-      "loss": 5.350383045536239e+28,
+      "loss": 0.0051,
       "step": 1950
     },
     {
       "epoch": 38.0,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.8017,
-      "eval_samples_per_second": 57.379,
-      "eval_steps_per_second": 7.484,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 1.2492655515670776,
+      "eval_runtime": 0.8062,
+      "eval_samples_per_second": 57.059,
+      "eval_steps_per_second": 7.443,
       "step": 1957
     },
     {
       "epoch": 38.06,
       "learning_rate": 2.2703818369453045e-06,
-      "loss": 4.755896040476657e+28,
+      "loss": 0.0228,
       "step": 1960
     },
     {
       "epoch": 38.25,
       "learning_rate": 1.9865841073271413e-06,
-      "loss": 6.985222498344749e+28,
+      "loss": 0.0322,
       "step": 1970
     },
     {
       "epoch": 38.45,
       "learning_rate": 1.7027863777089785e-06,
-      "loss": 5.499004607906475e+28,
+      "loss": 0.0503,
       "step": 1980
     },
     {
       "epoch": 38.64,
       "learning_rate": 1.4189886480908153e-06,
-      "loss": 6.242113930914931e+28,
+      "loss": 0.039,
       "step": 1990
     },
     {
       "epoch": 38.83,
       "learning_rate": 1.1351909184726523e-06,
-      "loss": 5.64762654806603e+28,
+      "loss": 0.0123,
       "step": 2000
     },
     {
       "epoch": 38.99,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7817,
-      "eval_samples_per_second": 58.848,
-      "eval_steps_per_second": 7.676,
+      "eval_accuracy": 0.7391304347826086,
+      "eval_loss": 1.2537580728530884,
+      "eval_runtime": 0.7792,
+      "eval_samples_per_second": 59.037,
+      "eval_steps_per_second": 7.7,
       "step": 2008
     },
     {
       "epoch": 39.03,
       "learning_rate": 8.513931888544892e-07,
-      "loss": 5.895329403875969e+28,
+      "loss": 0.0046,
       "step": 2010
     },
     {
       "epoch": 39.22,
       "learning_rate": 5.675954592363261e-07,
-      "loss": 6.687978618025639e+28,
+      "loss": 0.0218,
       "step": 2020
     },
     {
       "epoch": 39.42,
       "learning_rate": 2.8379772961816306e-07,
-      "loss": 6.093491612966058e+28,
+      "loss": 0.085,
       "step": 2030
     },
     {
       "epoch": 39.61,
       "learning_rate": 0.0,
-      "loss": 5.3503826677469204e+28,
+      "loss": 0.0546,
       "step": 2040
     },
     {
       "epoch": 39.61,
-      "eval_accuracy": 0.717391304347826,
-      "eval_loss": 2.584726139752747e+28,
-      "eval_runtime": 0.7909,
-      "eval_samples_per_second": 58.159,
-      "eval_steps_per_second": 7.586,
+      "eval_accuracy": 0.7608695652173914,
+      "eval_loss": 1.2530285120010376,
+      "eval_runtime": 0.8032,
+      "eval_samples_per_second": 57.272,
+      "eval_steps_per_second": 7.47,
       "step": 2040
     },
     {
       "epoch": 39.61,
       "step": 2040,
       "total_flos": 2.5142726714989363e+18,
-      "train_loss": 6.009952755433709e+28,
-      "train_runtime": 674.9098,
-      "train_samples_per_second": 48.54,
-      "train_steps_per_second": 3.023
+      "train_loss": 0.18828091986349546,
+      "train_runtime": 647.6496,
+      "train_samples_per_second": 50.583,
+      "train_steps_per_second": 3.15
     }
   ],
   "logging_steps": 10,