End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.8861386138613861
+      value: 0.8960396039603961
 ---
 
 <!-- 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 [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3726
-- Accuracy: 0.8861
+- Loss: 0.3583
+- Accuracy: 0.8960
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 19.65,
-    "eval_accuracy": 0.8787128712871287,
-    "eval_loss": 0.3504551351070404,
-    "eval_runtime": 3.6567,
-    "eval_samples_per_second": 110.481,
-    "eval_steps_per_second": 3.555,
+    "eval_accuracy": 0.8960396039603961,
+    "eval_loss": 0.3583451807498932,
+    "eval_runtime": 3.4101,
+    "eval_samples_per_second": 118.472,
+    "eval_steps_per_second": 3.812,
     "total_flos": 1.7752639958663823e+18,
-    "train_loss": 0.3740995813693319,
-    "train_runtime": 1232.3243,
-    "train_samples_per_second": 58.978,
-    "train_steps_per_second": 0.454
+    "train_loss": 0.37542668517146793,
+    "train_runtime": 934.7322,
+    "train_samples_per_second": 77.755,
+    "train_steps_per_second": 0.599
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 19.65,
-    "eval_accuracy": 0.8787128712871287,
-    "eval_loss": 0.3504551351070404,
-    "eval_runtime": 3.6567,
-    "eval_samples_per_second": 110.481,
-    "eval_steps_per_second": 3.555
+    "eval_accuracy": 0.8960396039603961,
+    "eval_loss": 0.3583451807498932,
+    "eval_runtime": 3.4101,
+    "eval_samples_per_second": 118.472,
+    "eval_steps_per_second": 3.812
 }

runs/Apr26_16-44-40_71f9cbb76c71/events.out.tfevents.1714150825.71f9cbb76c71.34.4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3a561d1f94c8f5833f88d9436db4f7f6af9bb3d28eec248b14fbb81bb5c7a02
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 19.65,
     "total_flos": 1.7752639958663823e+18,
-    "train_loss": 0.3740995813693319,
-    "train_runtime": 1232.3243,
-    "train_samples_per_second": 58.978,
-    "train_steps_per_second": 0.454
+    "train_loss": 0.37542668517146793,
+    "train_runtime": 934.7322,
+    "train_samples_per_second": 77.755,
+    "train_steps_per_second": 0.599
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.8787128712871287,
-  "best_model_checkpoint": "Electrcical-IMAGE-finetuned-eurosat/checkpoint-199",
+  "best_metric": 0.8960396039603961,
+  "best_model_checkpoint": "Electrcical-IMAGE-finetuned-eurosat/checkpoint-313",
   "epoch": 19.649122807017545,
   "eval_steps": 500,
   "global_step": 560,
@@ -10,584 +10,584 @@
   "log_history": [
     {
       "epoch": 0.35,
-      "grad_norm": 6.11484432220459,
+      "grad_norm": 10.917478561401367,
       "learning_rate": 8.92857142857143e-06,
-      "loss": 1.726,
+      "loss": 1.8574,
       "step": 10
     },
     {
       "epoch": 0.7,
-      "grad_norm": 4.320251941680908,
+      "grad_norm": 7.904279708862305,
       "learning_rate": 1.785714285714286e-05,
-      "loss": 1.5532,
+      "loss": 1.6143,
       "step": 20
     },
     {
       "epoch": 0.98,
-      "eval_accuracy": 0.6163366336633663,
-      "eval_loss": 1.170404314994812,
-      "eval_runtime": 5.3474,
-      "eval_samples_per_second": 75.551,
-      "eval_steps_per_second": 2.431,
+      "eval_accuracy": 0.5346534653465347,
+      "eval_loss": 1.2882142066955566,
+      "eval_runtime": 3.1902,
+      "eval_samples_per_second": 126.638,
+      "eval_steps_per_second": 4.075,
       "step": 28
     },
     {
       "epoch": 1.05,
-      "grad_norm": 4.363418102264404,
+      "grad_norm": 12.139655113220215,
       "learning_rate": 2.6785714285714288e-05,
-      "loss": 1.2686,
+      "loss": 1.3248,
       "step": 30
     },
     {
       "epoch": 1.4,
-      "grad_norm": 6.138254165649414,
+      "grad_norm": 6.75371789932251,
       "learning_rate": 3.571428571428572e-05,
-      "loss": 1.0232,
+      "loss": 1.1002,
       "step": 40
     },
     {
       "epoch": 1.75,
-      "grad_norm": 7.400917053222656,
+      "grad_norm": 16.99814224243164,
       "learning_rate": 4.464285714285715e-05,
-      "loss": 0.8115,
+      "loss": 0.8597,
       "step": 50
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.7673267326732673,
-      "eval_loss": 0.6826810836791992,
-      "eval_runtime": 3.1989,
-      "eval_samples_per_second": 126.295,
+      "eval_accuracy": 0.7648514851485149,
+      "eval_loss": 0.7302265763282776,
+      "eval_runtime": 3.1986,
+      "eval_samples_per_second": 126.304,
       "eval_steps_per_second": 4.064,
       "step": 57
     },
     {
       "epoch": 2.11,
-      "grad_norm": 7.412561893463135,
+      "grad_norm": 6.605724811553955,
       "learning_rate": 4.960317460317461e-05,
-      "loss": 0.6904,
+      "loss": 0.7264,
       "step": 60
     },
     {
       "epoch": 2.46,
-      "grad_norm": 8.475811004638672,
+      "grad_norm": 9.7291841506958,
       "learning_rate": 4.8611111111111115e-05,
-      "loss": 0.5908,
+      "loss": 0.6225,
       "step": 70
     },
     {
       "epoch": 2.81,
-      "grad_norm": 32.85942077636719,
+      "grad_norm": 7.020862579345703,
       "learning_rate": 4.761904761904762e-05,
-      "loss": 0.5513,
+      "loss": 0.5858,
       "step": 80
     },
     {
       "epoch": 2.98,
-      "eval_accuracy": 0.8415841584158416,
-      "eval_loss": 0.4524537920951843,
-      "eval_runtime": 3.2851,
-      "eval_samples_per_second": 122.978,
-      "eval_steps_per_second": 3.957,
+      "eval_accuracy": 0.8465346534653465,
+      "eval_loss": 0.48486289381980896,
+      "eval_runtime": 3.1164,
+      "eval_samples_per_second": 129.637,
+      "eval_steps_per_second": 4.171,
       "step": 85
     },
     {
       "epoch": 3.16,
-      "grad_norm": 5.583408355712891,
+      "grad_norm": 5.680540561676025,
       "learning_rate": 4.662698412698413e-05,
-      "loss": 0.5107,
+      "loss": 0.5214,
       "step": 90
     },
     {
       "epoch": 3.51,
-      "grad_norm": 7.957437038421631,
+      "grad_norm": 7.502087116241455,
       "learning_rate": 4.563492063492064e-05,
-      "loss": 0.4767,
+      "loss": 0.4695,
       "step": 100
     },
     {
       "epoch": 3.86,
-      "grad_norm": 6.342191219329834,
+      "grad_norm": 7.262702465057373,
       "learning_rate": 4.464285714285715e-05,
-      "loss": 0.455,
+      "loss": 0.4332,
       "step": 110
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8539603960396039,
-      "eval_loss": 0.40117621421813965,
-      "eval_runtime": 3.1407,
-      "eval_samples_per_second": 128.633,
-      "eval_steps_per_second": 4.139,
+      "eval_accuracy": 0.8613861386138614,
+      "eval_loss": 0.4274178743362427,
+      "eval_runtime": 3.1661,
+      "eval_samples_per_second": 127.603,
+      "eval_steps_per_second": 4.106,
       "step": 114
     },
     {
       "epoch": 4.21,
-      "grad_norm": 5.71556282043457,
+      "grad_norm": 6.890863418579102,
       "learning_rate": 4.3650793650793655e-05,
-      "loss": 0.4405,
+      "loss": 0.4297,
       "step": 120
     },
     {
       "epoch": 4.56,
-      "grad_norm": 6.5584306716918945,
+      "grad_norm": 6.25393533706665,
       "learning_rate": 4.265873015873016e-05,
-      "loss": 0.4179,
+      "loss": 0.4149,
       "step": 130
     },
     {
       "epoch": 4.91,
-      "grad_norm": 6.672788143157959,
+      "grad_norm": 7.106054782867432,
       "learning_rate": 4.166666666666667e-05,
-      "loss": 0.3901,
+      "loss": 0.4054,
       "step": 140
     },
     {
       "epoch": 4.98,
-      "eval_accuracy": 0.8613861386138614,
-      "eval_loss": 0.38241395354270935,
-      "eval_runtime": 3.2535,
-      "eval_samples_per_second": 124.173,
-      "eval_steps_per_second": 3.996,
+      "eval_accuracy": 0.8787128712871287,
+      "eval_loss": 0.3687044382095337,
+      "eval_runtime": 3.1309,
+      "eval_samples_per_second": 129.037,
+      "eval_steps_per_second": 4.152,
       "step": 142
     },
     {
       "epoch": 5.26,
-      "grad_norm": 8.260834693908691,
+      "grad_norm": 6.924484729766846,
       "learning_rate": 4.067460317460318e-05,
-      "loss": 0.3765,
+      "loss": 0.3727,
       "step": 150
     },
     {
       "epoch": 5.61,
-      "grad_norm": 9.854930877685547,
+      "grad_norm": 5.540406703948975,
       "learning_rate": 3.968253968253968e-05,
-      "loss": 0.3419,
+      "loss": 0.3299,
       "step": 160
     },
     {
       "epoch": 5.96,
-      "grad_norm": 8.068675994873047,
+      "grad_norm": 6.109258651733398,
       "learning_rate": 3.8690476190476195e-05,
-      "loss": 0.4042,
+      "loss": 0.3826,
       "step": 170
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8638613861386139,
-      "eval_loss": 0.37970858812332153,
-      "eval_runtime": 3.1263,
-      "eval_samples_per_second": 129.226,
-      "eval_steps_per_second": 4.158,
+      "eval_accuracy": 0.8613861386138614,
+      "eval_loss": 0.37880900502204895,
+      "eval_runtime": 3.1895,
+      "eval_samples_per_second": 126.667,
+      "eval_steps_per_second": 4.076,
       "step": 171
     },
     {
       "epoch": 6.32,
-      "grad_norm": 5.019181728363037,
+      "grad_norm": 4.945355415344238,
       "learning_rate": 3.76984126984127e-05,
-      "loss": 0.3335,
+      "loss": 0.3249,
       "step": 180
     },
     {
       "epoch": 6.67,
-      "grad_norm": 5.983124256134033,
+      "grad_norm": 4.232621669769287,
       "learning_rate": 3.6706349206349205e-05,
-      "loss": 0.3591,
+      "loss": 0.3561,
       "step": 190
     },
     {
       "epoch": 6.98,
-      "eval_accuracy": 0.8787128712871287,
-      "eval_loss": 0.3504551351070404,
-      "eval_runtime": 3.3207,
-      "eval_samples_per_second": 121.661,
-      "eval_steps_per_second": 3.915,
+      "eval_accuracy": 0.8935643564356436,
+      "eval_loss": 0.36995938420295715,
+      "eval_runtime": 3.1439,
+      "eval_samples_per_second": 128.503,
+      "eval_steps_per_second": 4.135,
       "step": 199
     },
     {
       "epoch": 7.02,
-      "grad_norm": 5.728125095367432,
+      "grad_norm": 5.742563247680664,
       "learning_rate": 3.571428571428572e-05,
-      "loss": 0.3308,
+      "loss": 0.3097,
       "step": 200
     },
     {
       "epoch": 7.37,
-      "grad_norm": 6.73638916015625,
+      "grad_norm": 6.4120402336120605,
       "learning_rate": 3.472222222222222e-05,
-      "loss": 0.3228,
+      "loss": 0.3167,
       "step": 210
     },
     {
       "epoch": 7.72,
-      "grad_norm": 9.998391151428223,
+      "grad_norm": 6.971658229827881,
       "learning_rate": 3.3730158730158734e-05,
-      "loss": 0.2989,
+      "loss": 0.2838,
       "step": 220
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8613861386138614,
-      "eval_loss": 0.355110764503479,
-      "eval_runtime": 3.1811,
-      "eval_samples_per_second": 127.002,
-      "eval_steps_per_second": 4.087,
+      "eval_accuracy": 0.8811881188118812,
+      "eval_loss": 0.3550195097923279,
+      "eval_runtime": 3.0809,
+      "eval_samples_per_second": 131.129,
+      "eval_steps_per_second": 4.22,
       "step": 228
     },
     {
       "epoch": 8.07,
-      "grad_norm": 6.521678924560547,
+      "grad_norm": 6.908318042755127,
       "learning_rate": 3.273809523809524e-05,
-      "loss": 0.2857,
+      "loss": 0.2828,
       "step": 230
     },
     {
       "epoch": 8.42,
-      "grad_norm": 6.126306056976318,
+      "grad_norm": 5.122198104858398,
       "learning_rate": 3.1746031746031745e-05,
-      "loss": 0.2748,
+      "loss": 0.267,
       "step": 240
     },
     {
       "epoch": 8.77,
-      "grad_norm": 7.768061637878418,
+      "grad_norm": 4.9148077964782715,
       "learning_rate": 3.075396825396826e-05,
-      "loss": 0.3029,
+      "loss": 0.2897,
       "step": 250
     },
     {
       "epoch": 8.98,
-      "eval_accuracy": 0.8663366336633663,
-      "eval_loss": 0.36247390508651733,
-      "eval_runtime": 3.1822,
-      "eval_samples_per_second": 126.958,
-      "eval_steps_per_second": 4.085,
+      "eval_accuracy": 0.8886138613861386,
+      "eval_loss": 0.3698354661464691,
+      "eval_runtime": 3.2226,
+      "eval_samples_per_second": 125.366,
+      "eval_steps_per_second": 4.034,
       "step": 256
     },
     {
       "epoch": 9.12,
-      "grad_norm": 6.541659355163574,
+      "grad_norm": 6.60095739364624,
       "learning_rate": 2.9761904761904762e-05,
-      "loss": 0.3138,
+      "loss": 0.3077,
       "step": 260
     },
     {
       "epoch": 9.47,
-      "grad_norm": 4.587341785430908,
+      "grad_norm": 5.082594394683838,
       "learning_rate": 2.876984126984127e-05,
-      "loss": 0.2556,
+      "loss": 0.245,
       "step": 270
     },
     {
       "epoch": 9.82,
-      "grad_norm": 3.2555856704711914,
+      "grad_norm": 3.717313766479492,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2606,
+      "loss": 0.2519,
       "step": 280
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.849009900990099,
-      "eval_loss": 0.3615105450153351,
-      "eval_runtime": 3.1469,
-      "eval_samples_per_second": 128.38,
-      "eval_steps_per_second": 4.131,
+      "eval_accuracy": 0.8836633663366337,
+      "eval_loss": 0.34588712453842163,
+      "eval_runtime": 3.2343,
+      "eval_samples_per_second": 124.913,
+      "eval_steps_per_second": 4.019,
       "step": 285
     },
     {
       "epoch": 10.18,
-      "grad_norm": 5.79010534286499,
+      "grad_norm": 4.068792819976807,
       "learning_rate": 2.6785714285714288e-05,
-      "loss": 0.2589,
+      "loss": 0.2456,
       "step": 290
     },
     {
       "epoch": 10.53,
-      "grad_norm": 7.236560821533203,
+      "grad_norm": 4.915719509124756,
       "learning_rate": 2.5793650793650796e-05,
-      "loss": 0.2615,
+      "loss": 0.2599,
       "step": 300
     },
     {
       "epoch": 10.88,
-      "grad_norm": 6.618343353271484,
+      "grad_norm": 4.958366870880127,
       "learning_rate": 2.4801587301587305e-05,
-      "loss": 0.2413,
+      "loss": 0.2194,
       "step": 310
     },
     {
       "epoch": 10.98,
-      "eval_accuracy": 0.8787128712871287,
-      "eval_loss": 0.34345176815986633,
-      "eval_runtime": 3.2537,
-      "eval_samples_per_second": 124.168,
-      "eval_steps_per_second": 3.996,
+      "eval_accuracy": 0.8960396039603961,
+      "eval_loss": 0.3583451807498932,
+      "eval_runtime": 3.1613,
+      "eval_samples_per_second": 127.797,
+      "eval_steps_per_second": 4.112,
       "step": 313
     },
     {
       "epoch": 11.23,
-      "grad_norm": 7.067264556884766,
+      "grad_norm": 8.63839054107666,
       "learning_rate": 2.380952380952381e-05,
-      "loss": 0.2329,
+      "loss": 0.2206,
       "step": 320
     },
     {
       "epoch": 11.58,
-      "grad_norm": 4.156674861907959,
+      "grad_norm": 6.567727565765381,
       "learning_rate": 2.281746031746032e-05,
-      "loss": 0.2286,
+      "loss": 0.22,
       "step": 330
     },
     {
       "epoch": 11.93,
-      "grad_norm": 6.207752227783203,
+      "grad_norm": 5.235540866851807,
       "learning_rate": 2.1825396825396827e-05,
-      "loss": 0.2051,
+      "loss": 0.1955,
       "step": 340
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8663366336633663,
-      "eval_loss": 0.3371087908744812,
-      "eval_runtime": 3.105,
-      "eval_samples_per_second": 130.112,
-      "eval_steps_per_second": 4.187,
+      "eval_accuracy": 0.8886138613861386,
+      "eval_loss": 0.3442213535308838,
+      "eval_runtime": 3.1157,
+      "eval_samples_per_second": 129.666,
+      "eval_steps_per_second": 4.172,
       "step": 342
     },
     {
       "epoch": 12.28,
-      "grad_norm": 4.795997142791748,
+      "grad_norm": 5.771058559417725,
       "learning_rate": 2.0833333333333336e-05,
-      "loss": 0.2052,
+      "loss": 0.1896,
       "step": 350
     },
     {
       "epoch": 12.63,
-      "grad_norm": 5.227004528045654,
+      "grad_norm": 4.845823287963867,
       "learning_rate": 1.984126984126984e-05,
-      "loss": 0.1962,
+      "loss": 0.1914,
       "step": 360
     },
     {
       "epoch": 12.98,
-      "grad_norm": 6.650285720825195,
+      "grad_norm": 6.198949813842773,
       "learning_rate": 1.884920634920635e-05,
-      "loss": 0.2477,
+      "loss": 0.2443,
       "step": 370
     },
     {
       "epoch": 12.98,
-      "eval_accuracy": 0.8638613861386139,
-      "eval_loss": 0.34505516290664673,
-      "eval_runtime": 3.309,
-      "eval_samples_per_second": 122.091,
-      "eval_steps_per_second": 3.929,
+      "eval_accuracy": 0.8787128712871287,
+      "eval_loss": 0.38009530305862427,
+      "eval_runtime": 3.1836,
+      "eval_samples_per_second": 126.902,
+      "eval_steps_per_second": 4.083,
       "step": 370
     },
     {
       "epoch": 13.33,
-      "grad_norm": 4.851602554321289,
+      "grad_norm": 5.5330023765563965,
       "learning_rate": 1.785714285714286e-05,
-      "loss": 0.2151,
+      "loss": 0.1938,
       "step": 380
     },
     {
       "epoch": 13.68,
-      "grad_norm": 6.44926118850708,
+      "grad_norm": 6.819355487823486,
       "learning_rate": 1.6865079365079367e-05,
-      "loss": 0.2271,
+      "loss": 0.207,
       "step": 390
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.8737623762376238,
-      "eval_loss": 0.3363986909389496,
-      "eval_runtime": 3.1022,
-      "eval_samples_per_second": 130.23,
-      "eval_steps_per_second": 4.191,
+      "eval_accuracy": 0.8861386138613861,
+      "eval_loss": 0.34993720054626465,
+      "eval_runtime": 3.1367,
+      "eval_samples_per_second": 128.798,
+      "eval_steps_per_second": 4.144,
       "step": 399
     },
     {
       "epoch": 14.04,
-      "grad_norm": 4.615256309509277,
+      "grad_norm": 3.808520555496216,
       "learning_rate": 1.5873015873015872e-05,
-      "loss": 0.2164,
+      "loss": 0.2176,
       "step": 400
     },
     {
       "epoch": 14.39,
-      "grad_norm": 4.279232025146484,
+      "grad_norm": 4.6104631423950195,
       "learning_rate": 1.4880952380952381e-05,
-      "loss": 0.1937,
+      "loss": 0.1725,
       "step": 410
     },
     {
       "epoch": 14.74,
-      "grad_norm": 5.792173385620117,
+      "grad_norm": 7.437196731567383,
       "learning_rate": 1.388888888888889e-05,
-      "loss": 0.2112,
+      "loss": 0.2078,
       "step": 420
     },
     {
       "epoch": 14.98,
-      "eval_accuracy": 0.8638613861386139,
-      "eval_loss": 0.35587263107299805,
-      "eval_runtime": 3.6554,
-      "eval_samples_per_second": 110.522,
-      "eval_steps_per_second": 3.556,
+      "eval_accuracy": 0.8836633663366337,
+      "eval_loss": 0.3701125383377075,
+      "eval_runtime": 3.2627,
+      "eval_samples_per_second": 123.824,
+      "eval_steps_per_second": 3.984,
       "step": 427
     },
     {
       "epoch": 15.09,
-      "grad_norm": 4.082373142242432,
+      "grad_norm": 4.747229099273682,
       "learning_rate": 1.2896825396825398e-05,
-      "loss": 0.1815,
+      "loss": 0.1775,
       "step": 430
     },
     {
       "epoch": 15.44,
-      "grad_norm": 6.473873138427734,
+      "grad_norm": 5.044639587402344,
       "learning_rate": 1.1904761904761905e-05,
-      "loss": 0.1884,
+      "loss": 0.1731,
       "step": 440
     },
     {
       "epoch": 15.79,
-      "grad_norm": 3.9221067428588867,
+      "grad_norm": 5.209843635559082,
       "learning_rate": 1.0912698412698414e-05,
-      "loss": 0.1902,
+      "loss": 0.1873,
       "step": 450
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.8737623762376238,
-      "eval_loss": 0.36296331882476807,
-      "eval_runtime": 3.2193,
-      "eval_samples_per_second": 125.494,
-      "eval_steps_per_second": 4.038,
+      "eval_accuracy": 0.8861386138613861,
+      "eval_loss": 0.37732651829719543,
+      "eval_runtime": 3.1869,
+      "eval_samples_per_second": 126.767,
+      "eval_steps_per_second": 4.079,
       "step": 456
     },
     {
       "epoch": 16.14,
-      "grad_norm": 3.29888916015625,
+      "grad_norm": 3.205880880355835,
       "learning_rate": 9.92063492063492e-06,
-      "loss": 0.1834,
+      "loss": 0.1728,
       "step": 460
     },
     {
       "epoch": 16.49,
-      "grad_norm": 5.7562432289123535,
+      "grad_norm": 5.715033054351807,
       "learning_rate": 8.92857142857143e-06,
-      "loss": 0.1797,
+      "loss": 0.1746,
       "step": 470
     },
     {
       "epoch": 16.84,
-      "grad_norm": 4.14888858795166,
+      "grad_norm": 3.9077258110046387,
       "learning_rate": 7.936507936507936e-06,
-      "loss": 0.1739,
+      "loss": 0.1697,
       "step": 480
     },
     {
       "epoch": 16.98,
-      "eval_accuracy": 0.8712871287128713,
-      "eval_loss": 0.36299142241477966,
-      "eval_runtime": 3.2007,
-      "eval_samples_per_second": 126.223,
-      "eval_steps_per_second": 4.062,
+      "eval_accuracy": 0.8861386138613861,
+      "eval_loss": 0.37531015276908875,
+      "eval_runtime": 3.1372,
+      "eval_samples_per_second": 128.778,
+      "eval_steps_per_second": 4.144,
       "step": 484
     },
     {
       "epoch": 17.19,
-      "grad_norm": 5.445935249328613,
+      "grad_norm": 6.2585344314575195,
       "learning_rate": 6.944444444444445e-06,
-      "loss": 0.1723,
+      "loss": 0.1774,
       "step": 490
     },
     {
       "epoch": 17.54,
-      "grad_norm": 10.137904167175293,
+      "grad_norm": 4.743368625640869,
       "learning_rate": 5.9523809523809525e-06,
-      "loss": 0.1721,
+      "loss": 0.1602,
       "step": 500
     },
     {
       "epoch": 17.89,
-      "grad_norm": 6.753917217254639,
+      "grad_norm": 5.987907409667969,
       "learning_rate": 4.96031746031746e-06,
-      "loss": 0.195,
+      "loss": 0.1812,
       "step": 510
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.8663366336633663,
-      "eval_loss": 0.3625456690788269,
-      "eval_runtime": 3.1415,
-      "eval_samples_per_second": 128.599,
-      "eval_steps_per_second": 4.138,
+      "eval_accuracy": 0.8910891089108911,
+      "eval_loss": 0.3747256398200989,
+      "eval_runtime": 3.3841,
+      "eval_samples_per_second": 119.38,
+      "eval_steps_per_second": 3.841,
       "step": 513
     },
     {
       "epoch": 18.25,
-      "grad_norm": 5.913514614105225,
+      "grad_norm": 4.54545259475708,
       "learning_rate": 3.968253968253968e-06,
-      "loss": 0.138,
+      "loss": 0.1375,
       "step": 520
     },
     {
       "epoch": 18.6,
-      "grad_norm": 5.631173610687256,
+      "grad_norm": 5.2062249183654785,
       "learning_rate": 2.9761904761904763e-06,
-      "loss": 0.1784,
+      "loss": 0.1713,
       "step": 530
     },
     {
       "epoch": 18.95,
-      "grad_norm": 3.3359873294830322,
+      "grad_norm": 5.484841823577881,
       "learning_rate": 1.984126984126984e-06,
-      "loss": 0.1621,
+      "loss": 0.151,
       "step": 540
     },
     {
       "epoch": 18.98,
-      "eval_accuracy": 0.8762376237623762,
-      "eval_loss": 0.35711079835891724,
-      "eval_runtime": 3.3137,
-      "eval_samples_per_second": 121.916,
-      "eval_steps_per_second": 3.923,
+      "eval_accuracy": 0.8861386138613861,
+      "eval_loss": 0.3735988438129425,
+      "eval_runtime": 3.2943,
+      "eval_samples_per_second": 122.636,
+      "eval_steps_per_second": 3.946,
       "step": 541
     },
     {
       "epoch": 19.3,
-      "grad_norm": 4.744650363922119,
+      "grad_norm": 5.083439350128174,
       "learning_rate": 9.92063492063492e-07,
-      "loss": 0.1726,
+      "loss": 0.163,
       "step": 550
     },
     {
       "epoch": 19.65,
-      "grad_norm": 5.039443492889404,
+      "grad_norm": 6.915902137756348,
       "learning_rate": 0.0,
-      "loss": 0.154,
+      "loss": 0.1567,
       "step": 560
     },
     {
       "epoch": 19.65,
-      "eval_accuracy": 0.8737623762376238,
-      "eval_loss": 0.3554615080356598,
-      "eval_runtime": 3.3898,
-      "eval_samples_per_second": 119.18,
-      "eval_steps_per_second": 3.835,
+      "eval_accuracy": 0.8861386138613861,
+      "eval_loss": 0.3726496696472168,
+      "eval_runtime": 3.3611,
+      "eval_samples_per_second": 120.199,
+      "eval_steps_per_second": 3.868,
       "step": 560
     },
     {
       "epoch": 19.65,
       "step": 560,
       "total_flos": 1.7752639958663823e+18,
-      "train_loss": 0.3740995813693319,
-      "train_runtime": 1232.3243,
-      "train_samples_per_second": 58.978,
-      "train_steps_per_second": 0.454
+      "train_loss": 0.37542668517146793,
+      "train_runtime": 934.7322,
+      "train_samples_per_second": 77.755,
+      "train_steps_per_second": 0.599
     }
   ],
   "logging_steps": 10,