End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.8737623762376238
+      value: 0.8787128712871287
 ---
 
 <!-- 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: 0.3555
-- Accuracy: 0.8738
+- Loss: 0.3505
+- Accuracy: 0.8787
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 19.649122807017545,
-    "eval_accuracy": 0.9034653465346535,
-    "eval_loss": 0.33881905674934387,
-    "eval_runtime": 4.2245,
-    "eval_samples_per_second": 95.632,
-    "eval_steps_per_second": 3.077,
+    "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,
     "total_flos": 1.7752639958663823e+18,
-    "train_loss": 0.17112446086747304,
-    "train_runtime": 1440.5794,
-    "train_samples_per_second": 50.452,
-    "train_steps_per_second": 0.389
+    "train_loss": 0.3740995813693319,
+    "train_runtime": 1232.3243,
+    "train_samples_per_second": 58.978,
+    "train_steps_per_second": 0.454
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 19.649122807017545,
-    "eval_accuracy": 0.9034653465346535,
-    "eval_loss": 0.33881905674934387,
-    "eval_runtime": 4.2245,
-    "eval_samples_per_second": 95.632,
-    "eval_steps_per_second": 3.077
+    "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
 }

runs/Apr26_16-15-29_71f9cbb76c71/events.out.tfevents.1714149387.71f9cbb76c71.34.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03d065adb021e431976c82aad08b304ba1c7af6328ce5f79056c9f24037356ba
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 19.649122807017545,
+    "epoch": 19.65,
     "total_flos": 1.7752639958663823e+18,
-    "train_loss": 0.17112446086747304,
-    "train_runtime": 1440.5794,
-    "train_samples_per_second": 50.452,
-    "train_steps_per_second": 0.389
+    "train_loss": 0.3740995813693319,
+    "train_runtime": 1232.3243,
+    "train_samples_per_second": 58.978,
+    "train_steps_per_second": 0.454
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.9034653465346535,
-  "best_model_checkpoint": "Electrcical-IMAGE-finetuned-eurosat/checkpoint-171",
+  "best_metric": 0.8787128712871287,
+  "best_model_checkpoint": "Electrcical-IMAGE-finetuned-eurosat/checkpoint-199",
   "epoch": 19.649122807017545,
   "eval_steps": 500,
   "global_step": 560,
@@ -9,585 +9,585 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.3508771929824561,
-      "grad_norm": 6.915553569793701,
+      "epoch": 0.35,
+      "grad_norm": 6.11484432220459,
       "learning_rate": 8.92857142857143e-06,
-      "loss": 0.2535,
+      "loss": 1.726,
       "step": 10
     },
     {
-      "epoch": 0.7017543859649122,
-      "grad_norm": 6.068195819854736,
+      "epoch": 0.7,
+      "grad_norm": 4.320251941680908,
       "learning_rate": 1.785714285714286e-05,
-      "loss": 0.2536,
+      "loss": 1.5532,
       "step": 20
     },
     {
-      "epoch": 0.9824561403508771,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.3073369264602661,
-      "eval_runtime": 4.5045,
-      "eval_samples_per_second": 89.687,
-      "eval_steps_per_second": 2.886,
+      "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,
       "step": 28
     },
     {
-      "epoch": 1.0526315789473684,
-      "grad_norm": 4.642627716064453,
+      "epoch": 1.05,
+      "grad_norm": 4.363418102264404,
       "learning_rate": 2.6785714285714288e-05,
-      "loss": 0.2506,
+      "loss": 1.2686,
       "step": 30
     },
     {
-      "epoch": 1.4035087719298245,
-      "grad_norm": 5.193023681640625,
+      "epoch": 1.4,
+      "grad_norm": 6.138254165649414,
       "learning_rate": 3.571428571428572e-05,
-      "loss": 0.2254,
+      "loss": 1.0232,
       "step": 40
     },
     {
-      "epoch": 1.7543859649122808,
-      "grad_norm": 6.378445148468018,
+      "epoch": 1.75,
+      "grad_norm": 7.400917053222656,
       "learning_rate": 4.464285714285715e-05,
-      "loss": 0.1702,
+      "loss": 0.8115,
       "step": 50
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8811881188118812,
-      "eval_loss": 0.32594048976898193,
-      "eval_runtime": 4.1579,
-      "eval_samples_per_second": 97.165,
-      "eval_steps_per_second": 3.127,
+      "eval_accuracy": 0.7673267326732673,
+      "eval_loss": 0.6826810836791992,
+      "eval_runtime": 3.1989,
+      "eval_samples_per_second": 126.295,
+      "eval_steps_per_second": 4.064,
       "step": 57
     },
     {
-      "epoch": 2.1052631578947367,
-      "grad_norm": 6.762551307678223,
+      "epoch": 2.11,
+      "grad_norm": 7.412561893463135,
       "learning_rate": 4.960317460317461e-05,
-      "loss": 0.2028,
+      "loss": 0.6904,
       "step": 60
     },
     {
-      "epoch": 2.456140350877193,
-      "grad_norm": 9.539087295532227,
+      "epoch": 2.46,
+      "grad_norm": 8.475811004638672,
       "learning_rate": 4.8611111111111115e-05,
-      "loss": 0.166,
+      "loss": 0.5908,
       "step": 70
     },
     {
-      "epoch": 2.807017543859649,
-      "grad_norm": 5.8399810791015625,
+      "epoch": 2.81,
+      "grad_norm": 32.85942077636719,
       "learning_rate": 4.761904761904762e-05,
-      "loss": 0.1528,
+      "loss": 0.5513,
       "step": 80
     },
     {
-      "epoch": 2.982456140350877,
-      "eval_accuracy": 0.8836633663366337,
-      "eval_loss": 0.3369888365268707,
-      "eval_runtime": 4.4971,
-      "eval_samples_per_second": 89.837,
-      "eval_steps_per_second": 2.891,
+      "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,
       "step": 85
     },
     {
-      "epoch": 3.1578947368421053,
-      "grad_norm": 2.9166860580444336,
+      "epoch": 3.16,
+      "grad_norm": 5.583408355712891,
       "learning_rate": 4.662698412698413e-05,
-      "loss": 0.1767,
+      "loss": 0.5107,
       "step": 90
     },
     {
-      "epoch": 3.5087719298245617,
-      "grad_norm": 3.477200984954834,
+      "epoch": 3.51,
+      "grad_norm": 7.957437038421631,
       "learning_rate": 4.563492063492064e-05,
-      "loss": 0.1722,
+      "loss": 0.4767,
       "step": 100
     },
     {
-      "epoch": 3.8596491228070176,
-      "grad_norm": 3.7534408569335938,
+      "epoch": 3.86,
+      "grad_norm": 6.342191219329834,
       "learning_rate": 4.464285714285715e-05,
-      "loss": 0.157,
+      "loss": 0.455,
       "step": 110
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8811881188118812,
-      "eval_loss": 0.3372877538204193,
-      "eval_runtime": 4.1256,
-      "eval_samples_per_second": 97.926,
-      "eval_steps_per_second": 3.151,
+      "eval_accuracy": 0.8539603960396039,
+      "eval_loss": 0.40117621421813965,
+      "eval_runtime": 3.1407,
+      "eval_samples_per_second": 128.633,
+      "eval_steps_per_second": 4.139,
       "step": 114
     },
     {
-      "epoch": 4.2105263157894735,
-      "grad_norm": 5.6130452156066895,
+      "epoch": 4.21,
+      "grad_norm": 5.71556282043457,
       "learning_rate": 4.3650793650793655e-05,
-      "loss": 0.1275,
+      "loss": 0.4405,
       "step": 120
     },
     {
-      "epoch": 4.56140350877193,
-      "grad_norm": 6.323825836181641,
+      "epoch": 4.56,
+      "grad_norm": 6.5584306716918945,
       "learning_rate": 4.265873015873016e-05,
-      "loss": 0.1678,
+      "loss": 0.4179,
       "step": 130
     },
     {
-      "epoch": 4.912280701754386,
-      "grad_norm": 8.424270629882812,
+      "epoch": 4.91,
+      "grad_norm": 6.672788143157959,
       "learning_rate": 4.166666666666667e-05,
-      "loss": 0.1577,
+      "loss": 0.3901,
       "step": 140
     },
     {
-      "epoch": 4.982456140350877,
-      "eval_accuracy": 0.8861386138613861,
-      "eval_loss": 0.3574221134185791,
-      "eval_runtime": 4.1778,
-      "eval_samples_per_second": 96.701,
-      "eval_steps_per_second": 3.112,
+      "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,
       "step": 142
     },
     {
-      "epoch": 5.2631578947368425,
-      "grad_norm": 7.5651936531066895,
+      "epoch": 5.26,
+      "grad_norm": 8.260834693908691,
       "learning_rate": 4.067460317460318e-05,
-      "loss": 0.1328,
+      "loss": 0.3765,
       "step": 150
     },
     {
-      "epoch": 5.614035087719298,
-      "grad_norm": 7.6120195388793945,
+      "epoch": 5.61,
+      "grad_norm": 9.854930877685547,
       "learning_rate": 3.968253968253968e-05,
-      "loss": 0.1428,
+      "loss": 0.3419,
       "step": 160
     },
     {
-      "epoch": 5.964912280701754,
-      "grad_norm": 9.945347785949707,
+      "epoch": 5.96,
+      "grad_norm": 8.068675994873047,
       "learning_rate": 3.8690476190476195e-05,
-      "loss": 0.1485,
+      "loss": 0.4042,
       "step": 170
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.9034653465346535,
-      "eval_loss": 0.33881905674934387,
-      "eval_runtime": 4.2693,
-      "eval_samples_per_second": 94.629,
-      "eval_steps_per_second": 3.045,
+      "eval_accuracy": 0.8638613861386139,
+      "eval_loss": 0.37970858812332153,
+      "eval_runtime": 3.1263,
+      "eval_samples_per_second": 129.226,
+      "eval_steps_per_second": 4.158,
       "step": 171
     },
     {
-      "epoch": 6.315789473684211,
-      "grad_norm": 5.8488969802856445,
+      "epoch": 6.32,
+      "grad_norm": 5.019181728363037,
       "learning_rate": 3.76984126984127e-05,
-      "loss": 0.1419,
+      "loss": 0.3335,
       "step": 180
     },
     {
-      "epoch": 6.666666666666667,
-      "grad_norm": 5.49446964263916,
+      "epoch": 6.67,
+      "grad_norm": 5.983124256134033,
       "learning_rate": 3.6706349206349205e-05,
-      "loss": 0.1715,
+      "loss": 0.3591,
       "step": 190
     },
     {
-      "epoch": 6.982456140350877,
-      "eval_accuracy": 0.8935643564356436,
-      "eval_loss": 0.3339557945728302,
-      "eval_runtime": 4.2324,
-      "eval_samples_per_second": 95.453,
-      "eval_steps_per_second": 3.072,
+      "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,
       "step": 199
     },
     {
-      "epoch": 7.017543859649122,
-      "grad_norm": 7.122160911560059,
+      "epoch": 7.02,
+      "grad_norm": 5.728125095367432,
       "learning_rate": 3.571428571428572e-05,
-      "loss": 0.2154,
+      "loss": 0.3308,
       "step": 200
     },
     {
-      "epoch": 7.368421052631579,
-      "grad_norm": 7.540256977081299,
+      "epoch": 7.37,
+      "grad_norm": 6.73638916015625,
       "learning_rate": 3.472222222222222e-05,
-      "loss": 0.2314,
+      "loss": 0.3228,
       "step": 210
     },
     {
-      "epoch": 7.719298245614035,
-      "grad_norm": 6.774203777313232,
+      "epoch": 7.72,
+      "grad_norm": 9.998391151428223,
       "learning_rate": 3.3730158730158734e-05,
-      "loss": 0.2079,
+      "loss": 0.2989,
       "step": 220
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8960396039603961,
-      "eval_loss": 0.3254563808441162,
-      "eval_runtime": 4.306,
-      "eval_samples_per_second": 93.823,
-      "eval_steps_per_second": 3.019,
+      "eval_accuracy": 0.8613861386138614,
+      "eval_loss": 0.355110764503479,
+      "eval_runtime": 3.1811,
+      "eval_samples_per_second": 127.002,
+      "eval_steps_per_second": 4.087,
       "step": 228
     },
     {
-      "epoch": 8.070175438596491,
-      "grad_norm": 8.224860191345215,
+      "epoch": 8.07,
+      "grad_norm": 6.521678924560547,
       "learning_rate": 3.273809523809524e-05,
-      "loss": 0.223,
+      "loss": 0.2857,
       "step": 230
     },
     {
-      "epoch": 8.421052631578947,
-      "grad_norm": 4.823063850402832,
+      "epoch": 8.42,
+      "grad_norm": 6.126306056976318,
       "learning_rate": 3.1746031746031745e-05,
-      "loss": 0.1889,
+      "loss": 0.2748,
       "step": 240
     },
     {
-      "epoch": 8.771929824561404,
-      "grad_norm": 6.559729099273682,
+      "epoch": 8.77,
+      "grad_norm": 7.768061637878418,
       "learning_rate": 3.075396825396826e-05,
-      "loss": 0.1846,
+      "loss": 0.3029,
       "step": 250
     },
     {
-      "epoch": 8.982456140350877,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.33677831292152405,
-      "eval_runtime": 4.1555,
-      "eval_samples_per_second": 97.221,
-      "eval_steps_per_second": 3.128,
+      "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,
       "step": 256
     },
     {
-      "epoch": 9.12280701754386,
-      "grad_norm": 4.693236827850342,
+      "epoch": 9.12,
+      "grad_norm": 6.541659355163574,
       "learning_rate": 2.9761904761904762e-05,
-      "loss": 0.1898,
+      "loss": 0.3138,
       "step": 260
     },
     {
-      "epoch": 9.473684210526315,
-      "grad_norm": 7.4912261962890625,
+      "epoch": 9.47,
+      "grad_norm": 4.587341785430908,
       "learning_rate": 2.876984126984127e-05,
-      "loss": 0.2097,
+      "loss": 0.2556,
       "step": 270
     },
     {
-      "epoch": 9.824561403508772,
-      "grad_norm": 5.2636213302612305,
+      "epoch": 9.82,
+      "grad_norm": 3.2555856704711914,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.1829,
+      "loss": 0.2606,
       "step": 280
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.317474901676178,
-      "eval_runtime": 4.3916,
-      "eval_samples_per_second": 91.994,
-      "eval_steps_per_second": 2.96,
+      "eval_accuracy": 0.849009900990099,
+      "eval_loss": 0.3615105450153351,
+      "eval_runtime": 3.1469,
+      "eval_samples_per_second": 128.38,
+      "eval_steps_per_second": 4.131,
       "step": 285
     },
     {
-      "epoch": 10.175438596491228,
-      "grad_norm": 4.9916558265686035,
+      "epoch": 10.18,
+      "grad_norm": 5.79010534286499,
       "learning_rate": 2.6785714285714288e-05,
-      "loss": 0.1684,
+      "loss": 0.2589,
       "step": 290
     },
     {
-      "epoch": 10.526315789473685,
-      "grad_norm": 6.921645641326904,
+      "epoch": 10.53,
+      "grad_norm": 7.236560821533203,
       "learning_rate": 2.5793650793650796e-05,
-      "loss": 0.1897,
+      "loss": 0.2615,
       "step": 300
     },
     {
-      "epoch": 10.87719298245614,
-      "grad_norm": 9.2024507522583,
+      "epoch": 10.88,
+      "grad_norm": 6.618343353271484,
       "learning_rate": 2.4801587301587305e-05,
-      "loss": 0.1874,
+      "loss": 0.2413,
       "step": 310
     },
     {
-      "epoch": 10.982456140350877,
-      "eval_accuracy": 0.8910891089108911,
-      "eval_loss": 0.3384435176849365,
-      "eval_runtime": 4.1888,
-      "eval_samples_per_second": 96.449,
-      "eval_steps_per_second": 3.104,
+      "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,
       "step": 313
     },
     {
-      "epoch": 11.228070175438596,
-      "grad_norm": 3.4831368923187256,
+      "epoch": 11.23,
+      "grad_norm": 7.067264556884766,
       "learning_rate": 2.380952380952381e-05,
-      "loss": 0.1468,
+      "loss": 0.2329,
       "step": 320
     },
     {
-      "epoch": 11.578947368421053,
-      "grad_norm": 30.23830795288086,
+      "epoch": 11.58,
+      "grad_norm": 4.156674861907959,
       "learning_rate": 2.281746031746032e-05,
-      "loss": 0.1723,
+      "loss": 0.2286,
       "step": 330
     },
     {
-      "epoch": 11.929824561403509,
-      "grad_norm": 6.060935974121094,
+      "epoch": 11.93,
+      "grad_norm": 6.207752227783203,
       "learning_rate": 2.1825396825396827e-05,
-      "loss": 0.1826,
+      "loss": 0.2051,
       "step": 340
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.32836616039276123,
-      "eval_runtime": 4.5565,
-      "eval_samples_per_second": 88.664,
-      "eval_steps_per_second": 2.853,
+      "eval_accuracy": 0.8663366336633663,
+      "eval_loss": 0.3371087908744812,
+      "eval_runtime": 3.105,
+      "eval_samples_per_second": 130.112,
+      "eval_steps_per_second": 4.187,
       "step": 342
     },
     {
-      "epoch": 12.280701754385966,
-      "grad_norm": 8.074810981750488,
+      "epoch": 12.28,
+      "grad_norm": 4.795997142791748,
       "learning_rate": 2.0833333333333336e-05,
-      "loss": 0.1798,
+      "loss": 0.2052,
       "step": 350
     },
     {
-      "epoch": 12.631578947368421,
-      "grad_norm": 6.300619125366211,
+      "epoch": 12.63,
+      "grad_norm": 5.227004528045654,
       "learning_rate": 1.984126984126984e-05,
-      "loss": 0.1516,
+      "loss": 0.1962,
       "step": 360
     },
     {
-      "epoch": 12.982456140350877,
-      "grad_norm": 8.272747993469238,
+      "epoch": 12.98,
+      "grad_norm": 6.650285720825195,
       "learning_rate": 1.884920634920635e-05,
-      "loss": 0.1911,
+      "loss": 0.2477,
       "step": 370
     },
     {
-      "epoch": 12.982456140350877,
-      "eval_accuracy": 0.8910891089108911,
-      "eval_loss": 0.3503906726837158,
-      "eval_runtime": 4.1749,
-      "eval_samples_per_second": 96.769,
-      "eval_steps_per_second": 3.114,
+      "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,
       "step": 370
     },
     {
-      "epoch": 13.333333333333334,
-      "grad_norm": 6.439492702484131,
+      "epoch": 13.33,
+      "grad_norm": 4.851602554321289,
       "learning_rate": 1.785714285714286e-05,
-      "loss": 0.1755,
+      "loss": 0.2151,
       "step": 380
     },
     {
-      "epoch": 13.68421052631579,
-      "grad_norm": 10.013567924499512,
+      "epoch": 13.68,
+      "grad_norm": 6.44926118850708,
       "learning_rate": 1.6865079365079367e-05,
-      "loss": 0.1552,
+      "loss": 0.2271,
       "step": 390
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.8935643564356436,
-      "eval_loss": 0.31156373023986816,
-      "eval_runtime": 4.4403,
-      "eval_samples_per_second": 90.985,
-      "eval_steps_per_second": 2.928,
+      "eval_accuracy": 0.8737623762376238,
+      "eval_loss": 0.3363986909389496,
+      "eval_runtime": 3.1022,
+      "eval_samples_per_second": 130.23,
+      "eval_steps_per_second": 4.191,
       "step": 399
     },
     {
-      "epoch": 14.035087719298245,
-      "grad_norm": 4.990965843200684,
+      "epoch": 14.04,
+      "grad_norm": 4.615256309509277,
       "learning_rate": 1.5873015873015872e-05,
-      "loss": 0.1737,
+      "loss": 0.2164,
       "step": 400
     },
     {
-      "epoch": 14.385964912280702,
-      "grad_norm": 5.88447380065918,
+      "epoch": 14.39,
+      "grad_norm": 4.279232025146484,
       "learning_rate": 1.4880952380952381e-05,
-      "loss": 0.1589,
+      "loss": 0.1937,
       "step": 410
     },
     {
-      "epoch": 14.736842105263158,
-      "grad_norm": 7.048079490661621,
+      "epoch": 14.74,
+      "grad_norm": 5.792173385620117,
       "learning_rate": 1.388888888888889e-05,
-      "loss": 0.1855,
+      "loss": 0.2112,
       "step": 420
     },
     {
-      "epoch": 14.982456140350877,
-      "eval_accuracy": 0.8960396039603961,
-      "eval_loss": 0.3131696581840515,
-      "eval_runtime": 4.1515,
-      "eval_samples_per_second": 97.315,
-      "eval_steps_per_second": 3.131,
+      "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,
       "step": 427
     },
     {
-      "epoch": 15.087719298245615,
-      "grad_norm": 5.223410129547119,
+      "epoch": 15.09,
+      "grad_norm": 4.082373142242432,
       "learning_rate": 1.2896825396825398e-05,
-      "loss": 0.1491,
+      "loss": 0.1815,
       "step": 430
     },
     {
-      "epoch": 15.43859649122807,
-      "grad_norm": 5.402275562286377,
+      "epoch": 15.44,
+      "grad_norm": 6.473873138427734,
       "learning_rate": 1.1904761904761905e-05,
-      "loss": 0.1627,
+      "loss": 0.1884,
       "step": 440
     },
     {
-      "epoch": 15.789473684210526,
-      "grad_norm": 8.536218643188477,
+      "epoch": 15.79,
+      "grad_norm": 3.9221067428588867,
       "learning_rate": 1.0912698412698414e-05,
-      "loss": 0.1537,
+      "loss": 0.1902,
       "step": 450
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.900990099009901,
-      "eval_loss": 0.30002227425575256,
-      "eval_runtime": 4.5205,
-      "eval_samples_per_second": 89.37,
-      "eval_steps_per_second": 2.876,
+      "eval_accuracy": 0.8737623762376238,
+      "eval_loss": 0.36296331882476807,
+      "eval_runtime": 3.2193,
+      "eval_samples_per_second": 125.494,
+      "eval_steps_per_second": 4.038,
       "step": 456
     },
     {
-      "epoch": 16.140350877192983,
-      "grad_norm": 6.947802543640137,
+      "epoch": 16.14,
+      "grad_norm": 3.29888916015625,
       "learning_rate": 9.92063492063492e-06,
-      "loss": 0.1511,
+      "loss": 0.1834,
       "step": 460
     },
     {
-      "epoch": 16.49122807017544,
-      "grad_norm": 2.9147555828094482,
+      "epoch": 16.49,
+      "grad_norm": 5.7562432289123535,
       "learning_rate": 8.92857142857143e-06,
-      "loss": 0.1232,
+      "loss": 0.1797,
       "step": 470
     },
     {
-      "epoch": 16.842105263157894,
-      "grad_norm": 4.058919906616211,
+      "epoch": 16.84,
+      "grad_norm": 4.14888858795166,
       "learning_rate": 7.936507936507936e-06,
-      "loss": 0.118,
+      "loss": 0.1739,
       "step": 480
     },
     {
-      "epoch": 16.982456140350877,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.326092928647995,
-      "eval_runtime": 4.138,
-      "eval_samples_per_second": 97.632,
-      "eval_steps_per_second": 3.142,
+      "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,
       "step": 484
     },
     {
-      "epoch": 17.19298245614035,
-      "grad_norm": 8.858957290649414,
+      "epoch": 17.19,
+      "grad_norm": 5.445935249328613,
       "learning_rate": 6.944444444444445e-06,
-      "loss": 0.1479,
+      "loss": 0.1723,
       "step": 490
     },
     {
-      "epoch": 17.54385964912281,
-      "grad_norm": 6.7740373611450195,
+      "epoch": 17.54,
+      "grad_norm": 10.137904167175293,
       "learning_rate": 5.9523809523809525e-06,
-      "loss": 0.1512,
+      "loss": 0.1721,
       "step": 500
     },
     {
-      "epoch": 17.894736842105264,
-      "grad_norm": 5.1415534019470215,
+      "epoch": 17.89,
+      "grad_norm": 6.753917217254639,
       "learning_rate": 4.96031746031746e-06,
-      "loss": 0.1483,
+      "loss": 0.195,
       "step": 510
     },
     {
       "epoch": 18.0,
-      "eval_accuracy": 0.8960396039603961,
-      "eval_loss": 0.31427323818206787,
-      "eval_runtime": 4.8063,
-      "eval_samples_per_second": 84.056,
-      "eval_steps_per_second": 2.705,
+      "eval_accuracy": 0.8663366336633663,
+      "eval_loss": 0.3625456690788269,
+      "eval_runtime": 3.1415,
+      "eval_samples_per_second": 128.599,
+      "eval_steps_per_second": 4.138,
       "step": 513
     },
     {
-      "epoch": 18.24561403508772,
-      "grad_norm": 5.084235668182373,
+      "epoch": 18.25,
+      "grad_norm": 5.913514614105225,
       "learning_rate": 3.968253968253968e-06,
-      "loss": 0.1392,
+      "loss": 0.138,
       "step": 520
     },
     {
-      "epoch": 18.596491228070175,
-      "grad_norm": 4.290576934814453,
+      "epoch": 18.6,
+      "grad_norm": 5.631173610687256,
       "learning_rate": 2.9761904761904763e-06,
-      "loss": 0.1191,
+      "loss": 0.1784,
       "step": 530
     },
     {
-      "epoch": 18.94736842105263,
-      "grad_norm": 5.261678218841553,
+      "epoch": 18.95,
+      "grad_norm": 3.3359873294830322,
       "learning_rate": 1.984126984126984e-06,
-      "loss": 0.1169,
+      "loss": 0.1621,
       "step": 540
     },
     {
-      "epoch": 18.982456140350877,
-      "eval_accuracy": 0.9034653465346535,
-      "eval_loss": 0.3101979196071625,
-      "eval_runtime": 4.2151,
-      "eval_samples_per_second": 95.846,
-      "eval_steps_per_second": 3.084,
+      "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,
       "step": 541
     },
     {
-      "epoch": 19.29824561403509,
-      "grad_norm": 5.632396221160889,
+      "epoch": 19.3,
+      "grad_norm": 4.744650363922119,
       "learning_rate": 9.92063492063492e-07,
-      "loss": 0.1521,
+      "loss": 0.1726,
       "step": 550
     },
     {
-      "epoch": 19.649122807017545,
-      "grad_norm": 6.291007995605469,
+      "epoch": 19.65,
+      "grad_norm": 5.039443492889404,
       "learning_rate": 0.0,
-      "loss": 0.1268,
+      "loss": 0.154,
       "step": 560
     },
     {
-      "epoch": 19.649122807017545,
-      "eval_accuracy": 0.8985148514851485,
-      "eval_loss": 0.3104223906993866,
-      "eval_runtime": 4.1976,
-      "eval_samples_per_second": 96.244,
-      "eval_steps_per_second": 3.097,
+      "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,
       "step": 560
     },
     {
-      "epoch": 19.649122807017545,
+      "epoch": 19.65,
       "step": 560,
       "total_flos": 1.7752639958663823e+18,
-      "train_loss": 0.17112446086747304,
-      "train_runtime": 1440.5794,
-      "train_samples_per_second": 50.452,
-      "train_steps_per_second": 0.389
+      "train_loss": 0.3740995813693319,
+      "train_runtime": 1232.3243,
+      "train_samples_per_second": 58.978,
+      "train_steps_per_second": 0.454
     }
   ],
   "logging_steps": 10,