End of training

README.md

@@ -17,8 +17,8 @@ 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 an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2748
-- Accuracy: 0.9148
+- Loss: 0.2736
+- Accuracy: 0.9162
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 9.98,
-    "eval_accuracy": 0.894109396914446,
-    "eval_loss": 0.303189218044281,
-    "eval_runtime": 41.2026,
-    "eval_samples_per_second": 69.219,
-    "eval_steps_per_second": 1.092,
+    "eval_accuracy": 0.9161991584852734,
+    "eval_loss": 0.2736373543739319,
+    "eval_runtime": 39.8217,
+    "eval_samples_per_second": 71.619,
+    "eval_steps_per_second": 1.13,
     "total_flos": 6.364199987970048e+18,
-    "train_loss": 0.48602974390983583,
-    "train_runtime": 5351.8475,
-    "train_samples_per_second": 47.954,
-    "train_steps_per_second": 0.187
+    "train_loss": 0.19688368451595306,
+    "train_runtime": 4893.4366,
+    "train_samples_per_second": 52.446,
+    "train_steps_per_second": 0.204
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.98,
-    "eval_accuracy": 0.894109396914446,
-    "eval_loss": 0.303189218044281,
-    "eval_runtime": 41.2026,
-    "eval_samples_per_second": 69.219,
-    "eval_steps_per_second": 1.092
+    "eval_accuracy": 0.9161991584852734,
+    "eval_loss": 0.2736373543739319,
+    "eval_runtime": 39.8217,
+    "eval_samples_per_second": 71.619,
+    "eval_steps_per_second": 1.13
 }

runs/May10_13-56-27_5ef4b5b162fa/events.out.tfevents.1715359858.5ef4b5b162fa.34.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b80d0ab69948e52e7db9bdefa3d711d596c039e7f7835b6241bf2df29d7040ae
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.98,
     "total_flos": 6.364199987970048e+18,
-    "train_loss": 0.48602974390983583,
-    "train_runtime": 5351.8475,
-    "train_samples_per_second": 47.954,
-    "train_steps_per_second": 0.187
+    "train_loss": 0.19688368451595306,
+    "train_runtime": 4893.4366,
+    "train_samples_per_second": 52.446,
+    "train_steps_per_second": 0.204
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.894109396914446,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-classification/checkpoint-1000",
+  "best_metric": 0.9161991584852734,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-classification/checkpoint-802",
   "epoch": 9.975062344139651,
   "eval_steps": 500,
   "global_step": 1000,
@@ -10,802 +10,802 @@
   "log_history": [
     {
       "epoch": 0.1,
-      "grad_norm": 5.313328742980957,
+      "grad_norm": 4.872883319854736,
       "learning_rate": 1e-05,
-      "loss": 2.0631,
+      "loss": 0.2252,
       "step": 10
     },
     {
       "epoch": 0.2,
-      "grad_norm": 5.775210857391357,
+      "grad_norm": 4.769281387329102,
       "learning_rate": 2e-05,
-      "loss": 1.7746,
+      "loss": 0.2281,
       "step": 20
     },
     {
       "epoch": 0.3,
-      "grad_norm": 4.816771984100342,
+      "grad_norm": 4.990223407745361,
       "learning_rate": 3e-05,
-      "loss": 1.5563,
+      "loss": 0.2281,
       "step": 30
     },
     {
       "epoch": 0.4,
-      "grad_norm": 6.007688999176025,
+      "grad_norm": 4.2269463539123535,
       "learning_rate": 4e-05,
-      "loss": 1.2361,
+      "loss": 0.2359,
       "step": 40
     },
     {
       "epoch": 0.5,
-      "grad_norm": 7.807107925415039,
+      "grad_norm": 5.273340702056885,
       "learning_rate": 5e-05,
-      "loss": 1.0685,
+      "loss": 0.2345,
       "step": 50
     },
     {
       "epoch": 0.6,
-      "grad_norm": 6.920202732086182,
+      "grad_norm": 6.166986465454102,
       "learning_rate": 6e-05,
-      "loss": 0.9716,
+      "loss": 0.2625,
       "step": 60
     },
     {
       "epoch": 0.7,
-      "grad_norm": 5.678104877471924,
+      "grad_norm": 5.066626071929932,
       "learning_rate": 7e-05,
-      "loss": 0.9274,
+      "loss": 0.2738,
       "step": 70
     },
     {
       "epoch": 0.8,
-      "grad_norm": 7.1128010749816895,
+      "grad_norm": 4.902966499328613,
       "learning_rate": 8e-05,
-      "loss": 0.8962,
+      "loss": 0.2793,
       "step": 80
     },
     {
       "epoch": 0.9,
-      "grad_norm": 8.700326919555664,
+      "grad_norm": 7.076936721801758,
       "learning_rate": 9e-05,
-      "loss": 0.8924,
+      "loss": 0.3018,
       "step": 90
     },
     {
       "epoch": 1.0,
-      "grad_norm": 7.496997833251953,
+      "grad_norm": 7.968544960021973,
       "learning_rate": 0.0001,
-      "loss": 0.8514,
+      "loss": 0.3193,
       "step": 100
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.705820476858345,
-      "eval_loss": 0.7862311005592346,
-      "eval_runtime": 38.0587,
-      "eval_samples_per_second": 74.937,
-      "eval_steps_per_second": 1.182,
+      "eval_accuracy": 0.8716690042075736,
+      "eval_loss": 0.3754436671733856,
+      "eval_runtime": 39.3279,
+      "eval_samples_per_second": 72.518,
+      "eval_steps_per_second": 1.144,
       "step": 100
     },
     {
       "epoch": 1.1,
-      "grad_norm": 5.842683792114258,
+      "grad_norm": 6.099853992462158,
       "learning_rate": 9.888888888888889e-05,
-      "loss": 0.7961,
+      "loss": 0.2997,
       "step": 110
     },
     {
       "epoch": 1.2,
-      "grad_norm": 7.082647323608398,
+      "grad_norm": 8.707666397094727,
       "learning_rate": 9.777777777777778e-05,
-      "loss": 0.8179,
+      "loss": 0.3066,
       "step": 120
     },
     {
       "epoch": 1.3,
-      "grad_norm": 6.531412601470947,
+      "grad_norm": 6.50078010559082,
       "learning_rate": 9.666666666666667e-05,
-      "loss": 0.772,
+      "loss": 0.2916,
       "step": 130
     },
     {
       "epoch": 1.4,
-      "grad_norm": 4.981847763061523,
+      "grad_norm": 6.648171901702881,
       "learning_rate": 9.555555555555557e-05,
-      "loss": 0.7298,
+      "loss": 0.318,
       "step": 140
     },
     {
       "epoch": 1.5,
-      "grad_norm": 6.9995198249816895,
+      "grad_norm": 5.3696746826171875,
       "learning_rate": 9.444444444444444e-05,
-      "loss": 0.6882,
+      "loss": 0.2995,
       "step": 150
     },
     {
       "epoch": 1.6,
-      "grad_norm": 4.9766693115234375,
+      "grad_norm": 4.172982692718506,
       "learning_rate": 9.333333333333334e-05,
-      "loss": 0.7048,
+      "loss": 0.2971,
       "step": 160
     },
     {
       "epoch": 1.7,
-      "grad_norm": 5.899988174438477,
+      "grad_norm": 5.951164245605469,
       "learning_rate": 9.222222222222223e-05,
-      "loss": 0.7254,
+      "loss": 0.3165,
       "step": 170
     },
     {
       "epoch": 1.8,
-      "grad_norm": 4.040374755859375,
+      "grad_norm": 5.611052513122559,
       "learning_rate": 9.111111111111112e-05,
-      "loss": 0.6543,
+      "loss": 0.2885,
       "step": 180
     },
     {
       "epoch": 1.9,
-      "grad_norm": 6.156548023223877,
+      "grad_norm": 6.186773777008057,
       "learning_rate": 9e-05,
-      "loss": 0.6329,
+      "loss": 0.2863,
       "step": 190
     },
     {
       "epoch": 2.0,
-      "grad_norm": 6.268504619598389,
+      "grad_norm": 8.462760925292969,
       "learning_rate": 8.888888888888889e-05,
-      "loss": 0.631,
+      "loss": 0.292,
       "step": 200
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.7885694249649369,
-      "eval_loss": 0.5491923093795776,
-      "eval_runtime": 38.032,
-      "eval_samples_per_second": 74.989,
-      "eval_steps_per_second": 1.183,
+      "eval_accuracy": 0.864656381486676,
+      "eval_loss": 0.4085911512374878,
+      "eval_runtime": 38.7867,
+      "eval_samples_per_second": 73.53,
+      "eval_steps_per_second": 1.16,
       "step": 200
     },
     {
       "epoch": 2.09,
-      "grad_norm": 6.425785541534424,
+      "grad_norm": 6.041274547576904,
       "learning_rate": 8.777777777777778e-05,
-      "loss": 0.612,
+      "loss": 0.2882,
       "step": 210
     },
     {
       "epoch": 2.19,
-      "grad_norm": 5.754735469818115,
+      "grad_norm": 5.336978435516357,
       "learning_rate": 8.666666666666667e-05,
-      "loss": 0.6117,
+      "loss": 0.269,
       "step": 220
     },
     {
       "epoch": 2.29,
-      "grad_norm": 5.51466703414917,
+      "grad_norm": 5.773736953735352,
       "learning_rate": 8.555555555555556e-05,
-      "loss": 0.6041,
+      "loss": 0.2675,
       "step": 230
     },
     {
       "epoch": 2.39,
-      "grad_norm": 5.170792102813721,
+      "grad_norm": 5.7196784019470215,
       "learning_rate": 8.444444444444444e-05,
-      "loss": 0.6057,
+      "loss": 0.256,
       "step": 240
     },
     {
       "epoch": 2.49,
-      "grad_norm": 5.319341659545898,
+      "grad_norm": 4.928184986114502,
       "learning_rate": 8.333333333333334e-05,
-      "loss": 0.556,
+      "loss": 0.2463,
       "step": 250
     },
     {
       "epoch": 2.59,
-      "grad_norm": 7.033453941345215,
+      "grad_norm": 7.451691627502441,
       "learning_rate": 8.222222222222222e-05,
-      "loss": 0.5918,
+      "loss": 0.2677,
       "step": 260
     },
     {
       "epoch": 2.69,
-      "grad_norm": 6.37441349029541,
+      "grad_norm": 5.755824089050293,
       "learning_rate": 8.111111111111112e-05,
-      "loss": 0.5603,
+      "loss": 0.2507,
       "step": 270
     },
     {
       "epoch": 2.79,
-      "grad_norm": 4.926444053649902,
+      "grad_norm": 5.09450101852417,
       "learning_rate": 8e-05,
-      "loss": 0.5827,
+      "loss": 0.2766,
       "step": 280
     },
     {
       "epoch": 2.89,
-      "grad_norm": 4.273934841156006,
+      "grad_norm": 5.240322589874268,
       "learning_rate": 7.88888888888889e-05,
-      "loss": 0.5406,
+      "loss": 0.2674,
       "step": 290
     },
     {
       "epoch": 2.99,
-      "grad_norm": 5.528937816619873,
+      "grad_norm": 4.865029811859131,
       "learning_rate": 7.777777777777778e-05,
-      "loss": 0.5516,
+      "loss": 0.2531,
       "step": 300
     },
     {
       "epoch": 2.99,
-      "eval_accuracy": 0.8078541374474053,
-      "eval_loss": 0.5194684863090515,
-      "eval_runtime": 38.7637,
-      "eval_samples_per_second": 73.574,
-      "eval_steps_per_second": 1.161,
+      "eval_accuracy": 0.894109396914446,
+      "eval_loss": 0.34070128202438354,
+      "eval_runtime": 39.4778,
+      "eval_samples_per_second": 72.243,
+      "eval_steps_per_second": 1.14,
       "step": 300
     },
     {
       "epoch": 3.09,
-      "grad_norm": 9.39556884765625,
+      "grad_norm": 4.403306007385254,
       "learning_rate": 7.666666666666667e-05,
-      "loss": 0.5039,
+      "loss": 0.2141,
       "step": 310
     },
     {
       "epoch": 3.19,
-      "grad_norm": 5.658567428588867,
+      "grad_norm": 4.91873836517334,
       "learning_rate": 7.555555555555556e-05,
-      "loss": 0.497,
+      "loss": 0.2257,
       "step": 320
     },
     {
       "epoch": 3.29,
-      "grad_norm": 5.974184513092041,
+      "grad_norm": 5.522466659545898,
       "learning_rate": 7.444444444444444e-05,
-      "loss": 0.5041,
+      "loss": 0.2419,
       "step": 330
     },
     {
       "epoch": 3.39,
-      "grad_norm": 5.876557350158691,
+      "grad_norm": 4.829797267913818,
       "learning_rate": 7.333333333333333e-05,
-      "loss": 0.4978,
+      "loss": 0.2522,
       "step": 340
     },
     {
       "epoch": 3.49,
-      "grad_norm": 7.005821704864502,
+      "grad_norm": 6.162938117980957,
       "learning_rate": 7.222222222222222e-05,
-      "loss": 0.5145,
+      "loss": 0.2533,
       "step": 350
     },
     {
       "epoch": 3.59,
-      "grad_norm": 4.967439651489258,
+      "grad_norm": 5.2231340408325195,
       "learning_rate": 7.111111111111112e-05,
-      "loss": 0.4803,
+      "loss": 0.2284,
       "step": 360
     },
     {
       "epoch": 3.69,
-      "grad_norm": 4.976831436157227,
+      "grad_norm": 4.9288716316223145,
       "learning_rate": 7e-05,
-      "loss": 0.4816,
+      "loss": 0.244,
       "step": 370
     },
     {
       "epoch": 3.79,
-      "grad_norm": 6.872544288635254,
+      "grad_norm": 4.194519996643066,
       "learning_rate": 6.88888888888889e-05,
-      "loss": 0.4811,
+      "loss": 0.2433,
       "step": 380
     },
     {
       "epoch": 3.89,
-      "grad_norm": 5.97269868850708,
+      "grad_norm": 5.996556282043457,
       "learning_rate": 6.777777777777778e-05,
-      "loss": 0.4978,
+      "loss": 0.2428,
       "step": 390
     },
     {
       "epoch": 3.99,
-      "grad_norm": 5.982359886169434,
+      "grad_norm": 5.954503536224365,
       "learning_rate": 6.666666666666667e-05,
-      "loss": 0.4597,
+      "loss": 0.2242,
       "step": 400
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8229312762973352,
-      "eval_loss": 0.48398053646087646,
-      "eval_runtime": 38.5586,
-      "eval_samples_per_second": 73.965,
-      "eval_steps_per_second": 1.167,
+      "eval_accuracy": 0.8927068723702665,
+      "eval_loss": 0.33171677589416504,
+      "eval_runtime": 39.2032,
+      "eval_samples_per_second": 72.749,
+      "eval_steps_per_second": 1.148,
       "step": 401
     },
     {
       "epoch": 4.09,
-      "grad_norm": 4.751348972320557,
+      "grad_norm": 5.060131072998047,
       "learning_rate": 6.555555555555556e-05,
-      "loss": 0.4443,
+      "loss": 0.209,
       "step": 410
     },
     {
       "epoch": 4.19,
-      "grad_norm": 4.549768447875977,
+      "grad_norm": 5.29101037979126,
       "learning_rate": 6.444444444444446e-05,
-      "loss": 0.423,
+      "loss": 0.2045,
       "step": 420
     },
     {
       "epoch": 4.29,
-      "grad_norm": 6.269441604614258,
+      "grad_norm": 7.005785942077637,
       "learning_rate": 6.333333333333333e-05,
-      "loss": 0.449,
+      "loss": 0.2063,
       "step": 430
     },
     {
       "epoch": 4.39,
-      "grad_norm": 4.581325531005859,
+      "grad_norm": 3.808072328567505,
       "learning_rate": 6.222222222222222e-05,
-      "loss": 0.4099,
+      "loss": 0.1935,
       "step": 440
     },
     {
       "epoch": 4.49,
-      "grad_norm": 5.31665563583374,
+      "grad_norm": 4.201041221618652,
       "learning_rate": 6.111111111111112e-05,
-      "loss": 0.4075,
+      "loss": 0.2062,
       "step": 450
     },
     {
       "epoch": 4.59,
-      "grad_norm": 5.4510416984558105,
+      "grad_norm": 5.4039835929870605,
       "learning_rate": 6e-05,
-      "loss": 0.4195,
+      "loss": 0.1949,
       "step": 460
     },
     {
       "epoch": 4.69,
-      "grad_norm": 5.312283515930176,
+      "grad_norm": 6.116047382354736,
       "learning_rate": 5.8888888888888896e-05,
-      "loss": 0.3829,
+      "loss": 0.2042,
       "step": 470
     },
     {
       "epoch": 4.79,
-      "grad_norm": 5.865201950073242,
+      "grad_norm": 4.799409866333008,
       "learning_rate": 5.7777777777777776e-05,
-      "loss": 0.4209,
+      "loss": 0.2135,
       "step": 480
     },
     {
       "epoch": 4.89,
-      "grad_norm": 5.123416423797607,
+      "grad_norm": 5.615736484527588,
       "learning_rate": 5.666666666666667e-05,
-      "loss": 0.4094,
+      "loss": 0.2098,
       "step": 490
     },
     {
       "epoch": 4.99,
-      "grad_norm": 4.805253982543945,
+      "grad_norm": 4.442112922668457,
       "learning_rate": 5.555555555555556e-05,
-      "loss": 0.4094,
+      "loss": 0.2181,
       "step": 500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.8597475455820477,
-      "eval_loss": 0.3945947587490082,
-      "eval_runtime": 38.9464,
-      "eval_samples_per_second": 73.229,
-      "eval_steps_per_second": 1.155,
+      "eval_accuracy": 0.8983169705469846,
+      "eval_loss": 0.3073332607746124,
+      "eval_runtime": 39.2249,
+      "eval_samples_per_second": 72.709,
+      "eval_steps_per_second": 1.147,
       "step": 501
     },
     {
       "epoch": 5.09,
-      "grad_norm": 4.796308517456055,
+      "grad_norm": 3.2966055870056152,
       "learning_rate": 5.4444444444444446e-05,
-      "loss": 0.3555,
+      "loss": 0.1809,
       "step": 510
     },
     {
       "epoch": 5.19,
-      "grad_norm": 4.460272789001465,
+      "grad_norm": 3.9496381282806396,
       "learning_rate": 5.333333333333333e-05,
-      "loss": 0.3719,
+      "loss": 0.1904,
       "step": 520
     },
     {
       "epoch": 5.29,
-      "grad_norm": 5.035310745239258,
+      "grad_norm": 4.369997978210449,
       "learning_rate": 5.222222222222223e-05,
-      "loss": 0.3785,
+      "loss": 0.1817,
       "step": 530
     },
     {
       "epoch": 5.39,
-      "grad_norm": 5.145148754119873,
+      "grad_norm": 4.793045520782471,
       "learning_rate": 5.111111111111111e-05,
-      "loss": 0.3881,
+      "loss": 0.1889,
       "step": 540
     },
     {
       "epoch": 5.49,
-      "grad_norm": 6.646022319793701,
+      "grad_norm": 3.8780226707458496,
       "learning_rate": 5e-05,
-      "loss": 0.3776,
+      "loss": 0.1713,
       "step": 550
     },
     {
       "epoch": 5.59,
-      "grad_norm": 6.039920330047607,
+      "grad_norm": 5.775850296020508,
       "learning_rate": 4.888888888888889e-05,
-      "loss": 0.3891,
+      "loss": 0.1916,
       "step": 560
     },
     {
       "epoch": 5.69,
-      "grad_norm": 4.1056623458862305,
+      "grad_norm": 4.524603843688965,
       "learning_rate": 4.7777777777777784e-05,
-      "loss": 0.3369,
+      "loss": 0.1771,
       "step": 570
     },
     {
       "epoch": 5.79,
-      "grad_norm": 5.633683204650879,
+      "grad_norm": 4.741029262542725,
       "learning_rate": 4.666666666666667e-05,
-      "loss": 0.3606,
+      "loss": 0.1722,
       "step": 580
     },
     {
       "epoch": 5.89,
-      "grad_norm": 5.873307704925537,
+      "grad_norm": 4.951327323913574,
       "learning_rate": 4.555555555555556e-05,
-      "loss": 0.3386,
+      "loss": 0.156,
       "step": 590
     },
     {
       "epoch": 5.99,
-      "grad_norm": 5.3684983253479,
+      "grad_norm": 3.7907869815826416,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.3289,
+      "loss": 0.1829,
       "step": 600
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8572931276297335,
-      "eval_loss": 0.3786664605140686,
-      "eval_runtime": 40.3623,
-      "eval_samples_per_second": 70.66,
-      "eval_steps_per_second": 1.115,
+      "eval_accuracy": 0.9032258064516129,
+      "eval_loss": 0.31035441160202026,
+      "eval_runtime": 38.6346,
+      "eval_samples_per_second": 73.82,
+      "eval_steps_per_second": 1.165,
       "step": 601
     },
     {
       "epoch": 6.08,
-      "grad_norm": 4.093196392059326,
+      "grad_norm": 4.384369850158691,
       "learning_rate": 4.3333333333333334e-05,
-      "loss": 0.3251,
+      "loss": 0.1656,
       "step": 610
     },
     {
       "epoch": 6.18,
-      "grad_norm": 4.739712715148926,
+      "grad_norm": 5.536717891693115,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.3029,
+      "loss": 0.1573,
       "step": 620
     },
     {
       "epoch": 6.28,
-      "grad_norm": 3.9042539596557617,
+      "grad_norm": 5.531068801879883,
       "learning_rate": 4.111111111111111e-05,
-      "loss": 0.3269,
+      "loss": 0.1585,
       "step": 630
     },
     {
       "epoch": 6.38,
-      "grad_norm": 5.695065498352051,
+      "grad_norm": 4.487759590148926,
       "learning_rate": 4e-05,
-      "loss": 0.3318,
+      "loss": 0.1618,
       "step": 640
     },
     {
       "epoch": 6.48,
-      "grad_norm": 4.814879894256592,
+      "grad_norm": 4.500083923339844,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 0.3087,
+      "loss": 0.1567,
       "step": 650
     },
     {
       "epoch": 6.58,
-      "grad_norm": 4.633285999298096,
+      "grad_norm": 4.1614298820495605,
       "learning_rate": 3.777777777777778e-05,
-      "loss": 0.3416,
+      "loss": 0.1548,
       "step": 660
     },
     {
       "epoch": 6.68,
-      "grad_norm": 4.862430095672607,
+      "grad_norm": 4.733586311340332,
       "learning_rate": 3.6666666666666666e-05,
-      "loss": 0.3063,
+      "loss": 0.1607,
       "step": 670
     },
     {
       "epoch": 6.78,
-      "grad_norm": 4.930345058441162,
+      "grad_norm": 5.126717567443848,
       "learning_rate": 3.555555555555556e-05,
-      "loss": 0.3225,
+      "loss": 0.1542,
       "step": 680
     },
     {
       "epoch": 6.88,
-      "grad_norm": 4.587200164794922,
+      "grad_norm": 3.771245002746582,
       "learning_rate": 3.444444444444445e-05,
-      "loss": 0.3139,
+      "loss": 0.1455,
       "step": 690
     },
     {
       "epoch": 6.98,
-      "grad_norm": 5.570033550262451,
+      "grad_norm": 5.217066287994385,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.3145,
+      "loss": 0.1426,
       "step": 700
     },
     {
       "epoch": 6.99,
-      "eval_accuracy": 0.8772791023842917,
-      "eval_loss": 0.3536257743835449,
-      "eval_runtime": 42.7073,
-      "eval_samples_per_second": 66.78,
-      "eval_steps_per_second": 1.054,
+      "eval_accuracy": 0.9049789621318373,
+      "eval_loss": 0.29764774441719055,
+      "eval_runtime": 39.4594,
+      "eval_samples_per_second": 72.277,
+      "eval_steps_per_second": 1.14,
       "step": 701
     },
     {
       "epoch": 7.08,
-      "grad_norm": 5.562831401824951,
+      "grad_norm": 6.066614627838135,
       "learning_rate": 3.222222222222223e-05,
-      "loss": 0.2998,
+      "loss": 0.1429,
       "step": 710
     },
     {
       "epoch": 7.18,
-      "grad_norm": 5.306057453155518,
+      "grad_norm": 3.363190174102783,
       "learning_rate": 3.111111111111111e-05,
-      "loss": 0.2807,
+      "loss": 0.1385,
       "step": 720
     },
     {
       "epoch": 7.28,
-      "grad_norm": 5.671096324920654,
+      "grad_norm": 5.055250644683838,
       "learning_rate": 3e-05,
-      "loss": 0.2794,
+      "loss": 0.1373,
       "step": 730
     },
     {
       "epoch": 7.38,
-      "grad_norm": 5.412108421325684,
+      "grad_norm": 5.0901384353637695,
       "learning_rate": 2.8888888888888888e-05,
-      "loss": 0.2814,
+      "loss": 0.1392,
       "step": 740
     },
     {
       "epoch": 7.48,
-      "grad_norm": 4.249139308929443,
+      "grad_norm": 3.7394046783447266,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2531,
+      "loss": 0.1235,
       "step": 750
     },
     {
       "epoch": 7.58,
-      "grad_norm": 5.714820861816406,
+      "grad_norm": 4.67645263671875,
       "learning_rate": 2.6666666666666667e-05,
-      "loss": 0.2712,
+      "loss": 0.1466,
       "step": 760
     },
     {
       "epoch": 7.68,
-      "grad_norm": 5.9426469802856445,
+      "grad_norm": 4.7584381103515625,
       "learning_rate": 2.5555555555555554e-05,
-      "loss": 0.2589,
+      "loss": 0.1228,
       "step": 770
     },
     {
       "epoch": 7.78,
-      "grad_norm": 4.787610054016113,
+      "grad_norm": 5.486098289489746,
       "learning_rate": 2.4444444444444445e-05,
-      "loss": 0.2745,
+      "loss": 0.1355,
       "step": 780
     },
     {
       "epoch": 7.88,
-      "grad_norm": 3.9021761417388916,
+      "grad_norm": 5.71726131439209,
       "learning_rate": 2.3333333333333336e-05,
-      "loss": 0.2728,
+      "loss": 0.1299,
       "step": 790
     },
     {
       "epoch": 7.98,
-      "grad_norm": 5.153603553771973,
+      "grad_norm": 4.222195625305176,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.2911,
+      "loss": 0.1472,
       "step": 800
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.885343618513324,
-      "eval_loss": 0.32486388087272644,
-      "eval_runtime": 42.9219,
-      "eval_samples_per_second": 66.446,
-      "eval_steps_per_second": 1.048,
+      "eval_accuracy": 0.9161991584852734,
+      "eval_loss": 0.2736373543739319,
+      "eval_runtime": 39.0364,
+      "eval_samples_per_second": 73.06,
+      "eval_steps_per_second": 1.153,
       "step": 802
     },
     {
       "epoch": 8.08,
-      "grad_norm": 5.644151210784912,
+      "grad_norm": 3.898097038269043,
       "learning_rate": 2.111111111111111e-05,
-      "loss": 0.2559,
+      "loss": 0.1214,
       "step": 810
     },
     {
       "epoch": 8.18,
-      "grad_norm": 4.023806095123291,
+      "grad_norm": 4.837826728820801,
       "learning_rate": 2e-05,
-      "loss": 0.2415,
+      "loss": 0.1176,
       "step": 820
     },
     {
       "epoch": 8.28,
-      "grad_norm": 6.214153289794922,
+      "grad_norm": 4.239289283752441,
       "learning_rate": 1.888888888888889e-05,
-      "loss": 0.2447,
+      "loss": 0.1191,
       "step": 830
     },
     {
       "epoch": 8.38,
-      "grad_norm": 3.903085231781006,
+      "grad_norm": 4.8459320068359375,
       "learning_rate": 1.777777777777778e-05,
-      "loss": 0.2602,
+      "loss": 0.1233,
       "step": 840
     },
     {
       "epoch": 8.48,
-      "grad_norm": 5.1828131675720215,
+      "grad_norm": 4.943118095397949,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.2529,
+      "loss": 0.1274,
       "step": 850
     },
     {
       "epoch": 8.58,
-      "grad_norm": 4.840915203094482,
+      "grad_norm": 3.3787436485290527,
       "learning_rate": 1.5555555555555555e-05,
-      "loss": 0.2302,
+      "loss": 0.1277,
       "step": 860
     },
     {
       "epoch": 8.68,
-      "grad_norm": 5.424919128417969,
+      "grad_norm": 4.009765148162842,
       "learning_rate": 1.4444444444444444e-05,
-      "loss": 0.246,
+      "loss": 0.1044,
       "step": 870
     },
     {
       "epoch": 8.78,
-      "grad_norm": 6.8232316970825195,
+      "grad_norm": 3.8422799110412598,
       "learning_rate": 1.3333333333333333e-05,
-      "loss": 0.2435,
+      "loss": 0.1195,
       "step": 880
     },
     {
       "epoch": 8.88,
-      "grad_norm": 4.856207847595215,
+      "grad_norm": 2.927946090698242,
       "learning_rate": 1.2222222222222222e-05,
-      "loss": 0.2376,
+      "loss": 0.1077,
       "step": 890
     },
     {
       "epoch": 8.98,
-      "grad_norm": 5.239475250244141,
+      "grad_norm": 5.300693988800049,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.2487,
+      "loss": 0.1314,
       "step": 900
     },
     {
       "epoch": 9.0,
-      "eval_accuracy": 0.8937587657784011,
-      "eval_loss": 0.30939263105392456,
-      "eval_runtime": 39.5102,
-      "eval_samples_per_second": 72.184,
-      "eval_steps_per_second": 1.139,
+      "eval_accuracy": 0.9133941093969145,
+      "eval_loss": 0.27998465299606323,
+      "eval_runtime": 39.0226,
+      "eval_samples_per_second": 73.086,
+      "eval_steps_per_second": 1.153,
       "step": 902
     },
     {
       "epoch": 9.08,
-      "grad_norm": 5.607595443725586,
+      "grad_norm": 4.2179107666015625,
       "learning_rate": 1e-05,
-      "loss": 0.2253,
+      "loss": 0.1172,
       "step": 910
     },
     {
       "epoch": 9.18,
-      "grad_norm": 3.9586544036865234,
+      "grad_norm": 4.46065092086792,
       "learning_rate": 8.88888888888889e-06,
-      "loss": 0.2478,
+      "loss": 0.1076,
       "step": 920
     },
     {
       "epoch": 9.28,
-      "grad_norm": 3.8814537525177,
+      "grad_norm": 3.4499449729919434,
       "learning_rate": 7.777777777777777e-06,
-      "loss": 0.2173,
+      "loss": 0.1103,
       "step": 930
     },
     {
       "epoch": 9.38,
-      "grad_norm": 7.456895351409912,
+      "grad_norm": 4.1871418952941895,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 0.2315,
+      "loss": 0.1041,
       "step": 940
     },
     {
       "epoch": 9.48,
-      "grad_norm": 4.543576717376709,
+      "grad_norm": 4.666812896728516,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.2281,
+      "loss": 0.1175,
       "step": 950
     },
     {
       "epoch": 9.58,
-      "grad_norm": 4.659777641296387,
+      "grad_norm": 3.4452712535858154,
       "learning_rate": 4.444444444444445e-06,
-      "loss": 0.2324,
+      "loss": 0.1143,
       "step": 960
     },
     {
       "epoch": 9.68,
-      "grad_norm": 4.048426151275635,
+      "grad_norm": 4.585890293121338,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 0.2132,
+      "loss": 0.1156,
       "step": 970
     },
     {
       "epoch": 9.78,
-      "grad_norm": 3.969958543777466,
+      "grad_norm": 3.672333240509033,
       "learning_rate": 2.2222222222222225e-06,
-      "loss": 0.2076,
+      "loss": 0.1157,
       "step": 980
     },
     {
       "epoch": 9.88,
-      "grad_norm": 5.621995449066162,
+      "grad_norm": 3.946244478225708,
       "learning_rate": 1.1111111111111112e-06,
-      "loss": 0.2262,
+      "loss": 0.1019,
       "step": 990
     },
     {
       "epoch": 9.98,
-      "grad_norm": 4.668450355529785,
+      "grad_norm": 2.6940207481384277,
       "learning_rate": 0.0,
-      "loss": 0.2291,
+      "loss": 0.1139,
       "step": 1000
     },
     {
       "epoch": 9.98,
-      "eval_accuracy": 0.894109396914446,
-      "eval_loss": 0.303189218044281,
-      "eval_runtime": 39.0907,
-      "eval_samples_per_second": 72.958,
-      "eval_steps_per_second": 1.151,
+      "eval_accuracy": 0.914796633941094,
+      "eval_loss": 0.2748475670814514,
+      "eval_runtime": 39.0773,
+      "eval_samples_per_second": 72.984,
+      "eval_steps_per_second": 1.152,
       "step": 1000
     },
     {
       "epoch": 9.98,
       "step": 1000,
       "total_flos": 6.364199987970048e+18,
-      "train_loss": 0.48602974390983583,
-      "train_runtime": 5351.8475,
-      "train_samples_per_second": 47.954,
-      "train_steps_per_second": 0.187
+      "train_loss": 0.19688368451595306,
+      "train_runtime": 4893.4366,
+      "train_samples_per_second": 52.446,
+      "train_steps_per_second": 0.204
     }
   ],
   "logging_steps": 10,