End of training

README.md

@@ -5,6 +5,8 @@ license: apache-2.0
 metrics:
 - accuracy
 tags:
+- image-classification
+- vision
 - generated_from_trainer
 model-index:
 - name: only-lora-beans-vit-base-patch16-224-in21k
@@ -16,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # only-lora-beans-vit-base-patch16-224-in21k
 
-This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the beans dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0983
+- Loss: 1.2664
 - Accuracy: 0.3308
 
 ## Model description

all_results.json

@@ -1,11 +1,13 @@
 {
     "epoch": 10.0,
-    "eval_runtime": 0.7922,
-    "eval_samples_per_second": 167.879,
-    "eval_steps_per_second": 21.458,
+    "eval_accuracy": 0.3308270676691729,
+    "eval_loss": 1.2663869857788086,
+    "eval_runtime": 0.9217,
+    "eval_samples_per_second": 144.296,
+    "eval_steps_per_second": 18.444,
     "total_flos": 8.400578669044531e+17,
-    "train_loss": 0.9780362775692573,
-    "train_runtime": 137.5902,
-    "train_samples_per_second": 75.151,
-    "train_steps_per_second": 9.448
+    "train_loss": 0.9618942783429072,
+    "train_runtime": 164.295,
+    "train_samples_per_second": 62.936,
+    "train_steps_per_second": 7.913
 }

eval_results.json

@@ -1,6 +1,8 @@
 {
     "epoch": 10.0,
-    "eval_runtime": 0.7922,
-    "eval_samples_per_second": 167.879,
-    "eval_steps_per_second": 21.458
+    "eval_accuracy": 0.3308270676691729,
+    "eval_loss": 1.2663869857788086,
+    "eval_runtime": 0.9217,
+    "eval_samples_per_second": 144.296,
+    "eval_steps_per_second": 18.444
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 10.0,
     "total_flos": 8.400578669044531e+17,
-    "train_loss": 0.9780362775692573,
-    "train_runtime": 137.5902,
-    "train_samples_per_second": 75.151,
-    "train_steps_per_second": 9.448
+    "train_loss": 0.9618942783429072,
+    "train_runtime": 164.295,
+    "train_samples_per_second": 62.936,
+    "train_steps_per_second": 7.913
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 0.3851391077041626,
+  "best_model_checkpoint": "./beans_outputs/checkpoint-130",
   "epoch": 10.0,
   "eval_steps": 500,
   "global_step": 1300,
@@ -10,992 +10,1012 @@
   "log_history": [
     {
       "epoch": 0.07692307692307693,
-      "grad_norm": 2.0707900524139404,
+      "grad_norm": 1.2493035793304443,
       "learning_rate": 0.004961538461538462,
-      "loss": 0.8561,
+      "loss": 0.8568,
       "step": 10
     },
     {
       "epoch": 0.15384615384615385,
-      "grad_norm": 0.9913710951805115,
+      "grad_norm": 1.4747363328933716,
       "learning_rate": 0.004923076923076923,
-      "loss": 0.6755,
+      "loss": 0.6606,
       "step": 20
     },
     {
       "epoch": 0.23076923076923078,
-      "grad_norm": 1.5771081447601318,
+      "grad_norm": 0.8188544511795044,
       "learning_rate": 0.004884615384615385,
-      "loss": 0.4032,
+      "loss": 0.3787,
       "step": 30
     },
     {
       "epoch": 0.3076923076923077,
-      "grad_norm": 1.9519121646881104,
+      "grad_norm": 1.718248724937439,
       "learning_rate": 0.004846153846153846,
-      "loss": 0.3903,
+      "loss": 0.3154,
       "step": 40
     },
     {
       "epoch": 0.38461538461538464,
-      "grad_norm": 2.557398796081543,
+      "grad_norm": 2.2010860443115234,
       "learning_rate": 0.004807692307692308,
-      "loss": 0.5472,
+      "loss": 0.4932,
       "step": 50
     },
     {
       "epoch": 0.46153846153846156,
-      "grad_norm": 2.4644172191619873,
+      "grad_norm": 2.927755832672119,
       "learning_rate": 0.0047692307692307695,
-      "loss": 0.3894,
+      "loss": 0.561,
       "step": 60
     },
     {
       "epoch": 0.5384615384615384,
-      "grad_norm": 1.2416399717330933,
+      "grad_norm": 1.2203412055969238,
       "learning_rate": 0.004730769230769231,
-      "loss": 0.1827,
+      "loss": 0.3309,
       "step": 70
     },
     {
       "epoch": 0.6153846153846154,
-      "grad_norm": 0.8515920639038086,
+      "grad_norm": 0.08795254677534103,
       "learning_rate": 0.004692307692307693,
-      "loss": 0.3148,
+      "loss": 0.3261,
       "step": 80
     },
     {
       "epoch": 0.6923076923076923,
-      "grad_norm": 0.227506622672081,
+      "grad_norm": 0.4609662890434265,
       "learning_rate": 0.004653846153846154,
-      "loss": 0.237,
+      "loss": 0.195,
       "step": 90
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 2.7295713424682617,
+      "grad_norm": 0.9046596884727478,
       "learning_rate": 0.004615384615384616,
-      "loss": 0.4501,
+      "loss": 0.3352,
       "step": 100
     },
     {
       "epoch": 0.8461538461538461,
-      "grad_norm": 1.2349276542663574,
+      "grad_norm": 2.5637238025665283,
       "learning_rate": 0.0045769230769230765,
-      "loss": 0.6825,
+      "loss": 0.5961,
       "step": 110
     },
     {
       "epoch": 0.9230769230769231,
-      "grad_norm": 3.311713933944702,
+      "grad_norm": 0.32293158769607544,
       "learning_rate": 0.004538461538461539,
-      "loss": 0.4085,
+      "loss": 0.236,
       "step": 120
     },
     {
       "epoch": 1.0,
-      "grad_norm": 0.20723728835582733,
+      "grad_norm": 7.484982967376709,
       "learning_rate": 0.0045000000000000005,
-      "loss": 0.1292,
+      "loss": 0.6323,
       "step": 130
     },
     {
       "epoch": 1.0,
-      "eval_runtime": 0.749,
-      "eval_samples_per_second": 177.568,
-      "eval_steps_per_second": 22.697,
+      "eval_accuracy": 0.8721804511278195,
+      "eval_loss": 0.3851391077041626,
+      "eval_runtime": 0.8932,
+      "eval_samples_per_second": 148.91,
+      "eval_steps_per_second": 19.034,
       "step": 130
     },
     {
       "epoch": 1.0769230769230769,
-      "grad_norm": 1.5258536338806152,
+      "grad_norm": 0.12631376087665558,
       "learning_rate": 0.004461538461538462,
-      "loss": 0.3389,
+      "loss": 0.747,
       "step": 140
     },
     {
       "epoch": 1.1538461538461537,
-      "grad_norm": 2.0288751125335693,
+      "grad_norm": 1.5336790084838867,
       "learning_rate": 0.004423076923076923,
-      "loss": 0.2397,
+      "loss": 0.3062,
       "step": 150
     },
     {
       "epoch": 1.2307692307692308,
-      "grad_norm": 0.03801692649722099,
+      "grad_norm": 0.7460920214653015,
       "learning_rate": 0.004384615384615384,
-      "loss": 0.162,
+      "loss": 0.2088,
       "step": 160
     },
     {
       "epoch": 1.3076923076923077,
-      "grad_norm": 0.1472761183977127,
+      "grad_norm": 0.6954104900360107,
       "learning_rate": 0.004346153846153846,
-      "loss": 0.2687,
+      "loss": 0.2942,
       "step": 170
     },
     {
       "epoch": 1.3846153846153846,
-      "grad_norm": 4.054047107696533,
+      "grad_norm": 2.2199032306671143,
       "learning_rate": 0.004307692307692308,
-      "loss": 0.6044,
+      "loss": 0.315,
       "step": 180
     },
     {
       "epoch": 1.4615384615384617,
-      "grad_norm": 1.6007931232452393,
+      "grad_norm": 2.654019832611084,
       "learning_rate": 0.004269230769230769,
-      "loss": 0.6124,
+      "loss": 0.4203,
       "step": 190
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 0.15122832357883453,
+      "grad_norm": 1.460809350013733,
       "learning_rate": 0.004230769230769231,
-      "loss": 0.4298,
+      "loss": 0.4812,
       "step": 200
     },
     {
       "epoch": 1.6153846153846154,
-      "grad_norm": 1.2161375284194946,
+      "grad_norm": 0.9371921420097351,
       "learning_rate": 0.004192307692307692,
-      "loss": 0.5844,
+      "loss": 0.3436,
       "step": 210
     },
     {
       "epoch": 1.6923076923076923,
-      "grad_norm": 7.478526592254639,
+      "grad_norm": 0.573295533657074,
       "learning_rate": 0.004153846153846154,
-      "loss": 1.2806,
+      "loss": 0.7394,
       "step": 220
     },
     {
       "epoch": 1.7692307692307692,
-      "grad_norm": 1.2232049703598022,
+      "grad_norm": 3.2005550861358643,
       "learning_rate": 0.004115384615384615,
-      "loss": 0.9687,
+      "loss": 0.5694,
       "step": 230
     },
     {
       "epoch": 1.8461538461538463,
-      "grad_norm": 1.4465281963348389,
+      "grad_norm": 1.0854452848434448,
       "learning_rate": 0.004076923076923077,
-      "loss": 0.4883,
+      "loss": 0.6043,
       "step": 240
     },
     {
       "epoch": 1.9230769230769231,
-      "grad_norm": 3.471082925796509,
+      "grad_norm": 1.3743146657943726,
       "learning_rate": 0.0040384615384615385,
-      "loss": 0.6709,
+      "loss": 0.4406,
       "step": 250
     },
     {
       "epoch": 2.0,
-      "grad_norm": 2.699213981628418,
+      "grad_norm": 0.11435157805681229,
       "learning_rate": 0.004,
-      "loss": 0.6404,
+      "loss": 0.4102,
       "step": 260
     },
     {
       "epoch": 2.0,
-      "eval_runtime": 0.747,
-      "eval_samples_per_second": 178.056,
-      "eval_steps_per_second": 22.759,
+      "eval_accuracy": 0.6240601503759399,
+      "eval_loss": 1.1071540117263794,
+      "eval_runtime": 0.9165,
+      "eval_samples_per_second": 145.121,
+      "eval_steps_per_second": 18.549,
       "step": 260
     },
     {
       "epoch": 2.076923076923077,
-      "grad_norm": 1.1869909763336182,
+      "grad_norm": 0.9636590480804443,
       "learning_rate": 0.003961538461538462,
-      "loss": 0.6565,
+      "loss": 0.4669,
       "step": 270
     },
     {
       "epoch": 2.1538461538461537,
-      "grad_norm": 2.2662851810455322,
+      "grad_norm": 1.4855515956878662,
       "learning_rate": 0.003923076923076923,
-      "loss": 0.3753,
+      "loss": 0.4618,
       "step": 280
     },
     {
       "epoch": 2.230769230769231,
-      "grad_norm": 3.587059736251831,
+      "grad_norm": 1.6462602615356445,
       "learning_rate": 0.003884615384615385,
-      "loss": 0.9706,
+      "loss": 0.5266,
       "step": 290
     },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 1.530928611755371,
+      "grad_norm": 2.7035958766937256,
       "learning_rate": 0.0038461538461538464,
-      "loss": 1.2593,
+      "loss": 0.6045,
       "step": 300
     },
     {
       "epoch": 2.3846153846153846,
-      "grad_norm": 1.0865129232406616,
+      "grad_norm": 1.7253210544586182,
       "learning_rate": 0.0038076923076923075,
-      "loss": 0.9166,
+      "loss": 0.4711,
       "step": 310
     },
     {
       "epoch": 2.4615384615384617,
-      "grad_norm": 2.3256101608276367,
+      "grad_norm": 0.6617446541786194,
       "learning_rate": 0.003769230769230769,
-      "loss": 1.1882,
+      "loss": 0.5785,
       "step": 320
     },
     {
       "epoch": 2.5384615384615383,
-      "grad_norm": 0.5831055045127869,
+      "grad_norm": 0.6448426842689514,
       "learning_rate": 0.003730769230769231,
-      "loss": 1.2764,
+      "loss": 0.7907,
       "step": 330
     },
     {
       "epoch": 2.6153846153846154,
-      "grad_norm": 0.9471529126167297,
+      "grad_norm": 3.4137613773345947,
       "learning_rate": 0.0036923076923076927,
-      "loss": 1.349,
+      "loss": 0.9229,
       "step": 340
     },
     {
       "epoch": 2.6923076923076925,
-      "grad_norm": 0.6033434867858887,
+      "grad_norm": 5.124917507171631,
       "learning_rate": 0.003653846153846154,
-      "loss": 1.1626,
+      "loss": 1.0287,
       "step": 350
     },
     {
       "epoch": 2.769230769230769,
-      "grad_norm": 0.322643518447876,
+      "grad_norm": 1.2190972566604614,
       "learning_rate": 0.0036153846153846154,
-      "loss": 1.1705,
+      "loss": 0.7791,
       "step": 360
     },
     {
       "epoch": 2.8461538461538463,
-      "grad_norm": 0.6473853588104248,
+      "grad_norm": 0.6898893713951111,
       "learning_rate": 0.003576923076923077,
-      "loss": 1.1492,
+      "loss": 1.4277,
       "step": 370
     },
     {
       "epoch": 2.9230769230769234,
-      "grad_norm": 0.4563555121421814,
+      "grad_norm": 2.563831090927124,
       "learning_rate": 0.003538461538461539,
-      "loss": 1.1009,
+      "loss": 1.3955,
       "step": 380
     },
     {
       "epoch": 3.0,
-      "grad_norm": 0.6773648858070374,
+      "grad_norm": 2.519380569458008,
       "learning_rate": 0.0034999999999999996,
-      "loss": 1.1322,
+      "loss": 0.8964,
       "step": 390
     },
     {
       "epoch": 3.0,
-      "eval_runtime": 0.7759,
-      "eval_samples_per_second": 171.41,
-      "eval_steps_per_second": 21.91,
+      "eval_accuracy": 0.6616541353383458,
+      "eval_loss": 0.8284351825714111,
+      "eval_runtime": 0.9347,
+      "eval_samples_per_second": 142.291,
+      "eval_steps_per_second": 18.188,
       "step": 390
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 0.499421089887619,
+      "grad_norm": 4.3924150466918945,
       "learning_rate": 0.0034615384615384616,
-      "loss": 1.0926,
+      "loss": 1.2627,
       "step": 400
     },
     {
       "epoch": 3.1538461538461537,
-      "grad_norm": 0.20487019419670105,
+      "grad_norm": 1.773996114730835,
       "learning_rate": 0.003423076923076923,
-      "loss": 1.129,
+      "loss": 1.3465,
       "step": 410
     },
     {
       "epoch": 3.230769230769231,
-      "grad_norm": 0.5652111172676086,
+      "grad_norm": 3.1790945529937744,
       "learning_rate": 0.003384615384615385,
-      "loss": 1.0965,
+      "loss": 1.2822,
       "step": 420
     },
     {
       "epoch": 3.3076923076923075,
-      "grad_norm": 0.5159160494804382,
+      "grad_norm": 5.48622465133667,
       "learning_rate": 0.003346153846153846,
-      "loss": 1.0865,
+      "loss": 1.0807,
       "step": 430
     },
     {
       "epoch": 3.3846153846153846,
-      "grad_norm": 0.5972450971603394,
+      "grad_norm": 3.6509623527526855,
       "learning_rate": 0.0033076923076923075,
-      "loss": 1.1468,
+      "loss": 1.1806,
       "step": 440
     },
     {
       "epoch": 3.4615384615384617,
-      "grad_norm": 0.4497612714767456,
+      "grad_norm": 2.3183326721191406,
       "learning_rate": 0.0032692307692307695,
-      "loss": 1.0972,
+      "loss": 1.2542,
       "step": 450
     },
     {
       "epoch": 3.5384615384615383,
-      "grad_norm": 0.8494599461555481,
+      "grad_norm": 3.2686471939086914,
       "learning_rate": 0.003230769230769231,
-      "loss": 1.1196,
+      "loss": 1.2217,
       "step": 460
     },
     {
       "epoch": 3.6153846153846154,
-      "grad_norm": 0.21025516092777252,
+      "grad_norm": 0.33250316977500916,
       "learning_rate": 0.003192307692307692,
-      "loss": 1.1336,
+      "loss": 1.2107,
       "step": 470
     },
     {
       "epoch": 3.6923076923076925,
-      "grad_norm": 0.5746878385543823,
+      "grad_norm": 2.4644064903259277,
       "learning_rate": 0.0031538461538461538,
-      "loss": 1.0988,
+      "loss": 1.2287,
       "step": 480
     },
     {
       "epoch": 3.769230769230769,
-      "grad_norm": 0.20434798300266266,
+      "grad_norm": 0.9661989808082581,
       "learning_rate": 0.0031153846153846153,
-      "loss": 1.1101,
+      "loss": 1.1392,
       "step": 490
     },
     {
       "epoch": 3.8461538461538463,
-      "grad_norm": 0.15725131332874298,
+      "grad_norm": 1.0765918493270874,
       "learning_rate": 0.0030769230769230774,
-      "loss": 1.1133,
+      "loss": 1.1576,
       "step": 500
     },
     {
       "epoch": 3.9230769230769234,
-      "grad_norm": 0.12141856551170349,
+      "grad_norm": 0.5606147646903992,
       "learning_rate": 0.0030384615384615385,
-      "loss": 1.0987,
+      "loss": 1.1148,
       "step": 510
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.1281518936157227,
+      "grad_norm": 3.787473678588867,
       "learning_rate": 0.003,
-      "loss": 1.1021,
+      "loss": 1.155,
       "step": 520
     },
     {
       "epoch": 4.0,
-      "eval_runtime": 0.7689,
-      "eval_samples_per_second": 172.973,
-      "eval_steps_per_second": 22.109,
+      "eval_accuracy": 0.3383458646616541,
+      "eval_loss": 1.1070177555084229,
+      "eval_runtime": 0.9062,
+      "eval_samples_per_second": 146.767,
+      "eval_steps_per_second": 18.76,
       "step": 520
     },
     {
       "epoch": 4.076923076923077,
-      "grad_norm": 0.5692263841629028,
+      "grad_norm": 1.8376567363739014,
       "learning_rate": 0.0029615384615384616,
-      "loss": 1.1032,
+      "loss": 1.16,
       "step": 530
     },
     {
       "epoch": 4.153846153846154,
-      "grad_norm": 0.21811030805110931,
+      "grad_norm": 0.1573391705751419,
       "learning_rate": 0.002923076923076923,
-      "loss": 1.096,
+      "loss": 1.1133,
       "step": 540
     },
     {
       "epoch": 4.230769230769231,
-      "grad_norm": 0.44653287529945374,
+      "grad_norm": 1.7338229417800903,
       "learning_rate": 0.0028846153846153843,
-      "loss": 1.0986,
+      "loss": 1.1604,
       "step": 550
     },
     {
       "epoch": 4.3076923076923075,
-      "grad_norm": 0.4496428966522217,
+      "grad_norm": 1.532076358795166,
       "learning_rate": 0.002846153846153846,
-      "loss": 1.1244,
+      "loss": 1.171,
       "step": 560
     },
     {
       "epoch": 4.384615384615385,
-      "grad_norm": 0.34072381258010864,
+      "grad_norm": 1.2591376304626465,
       "learning_rate": 0.002807692307692308,
-      "loss": 1.0982,
+      "loss": 1.116,
       "step": 570
     },
     {
       "epoch": 4.461538461538462,
-      "grad_norm": 0.617884635925293,
+      "grad_norm": 1.7697715759277344,
       "learning_rate": 0.0027692307692307695,
-      "loss": 1.1137,
+      "loss": 1.1409,
       "step": 580
     },
     {
       "epoch": 4.538461538461538,
-      "grad_norm": 0.3251841068267822,
+      "grad_norm": 1.0740739107131958,
       "learning_rate": 0.0027307692307692306,
-      "loss": 1.1054,
+      "loss": 1.1458,
       "step": 590
     },
     {
       "epoch": 4.615384615384615,
-      "grad_norm": 0.11284901946783066,
+      "grad_norm": 0.46824851632118225,
       "learning_rate": 0.002692307692307692,
-      "loss": 1.0981,
+      "loss": 1.1056,
       "step": 600
     },
     {
       "epoch": 4.6923076923076925,
-      "grad_norm": 0.802207887172699,
+      "grad_norm": 1.8973933458328247,
       "learning_rate": 0.0026538461538461538,
-      "loss": 1.1134,
+      "loss": 1.1024,
       "step": 610
     },
     {
       "epoch": 4.769230769230769,
-      "grad_norm": 0.22614385187625885,
+      "grad_norm": 0.06765652447938919,
       "learning_rate": 0.0026153846153846158,
-      "loss": 1.1155,
+      "loss": 1.1514,
       "step": 620
     },
     {
       "epoch": 4.846153846153846,
-      "grad_norm": 0.3756657838821411,
+      "grad_norm": 0.8284403085708618,
       "learning_rate": 0.002576923076923077,
-      "loss": 1.1113,
+      "loss": 1.1799,
       "step": 630
     },
     {
       "epoch": 4.923076923076923,
-      "grad_norm": 0.6828984618186951,
+      "grad_norm": 1.611846923828125,
       "learning_rate": 0.0025384615384615385,
-      "loss": 1.1249,
+      "loss": 1.169,
       "step": 640
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1.2701197862625122,
+      "grad_norm": 3.048553705215454,
       "learning_rate": 0.0025,
-      "loss": 1.102,
+      "loss": 1.1069,
       "step": 650
     },
     {
       "epoch": 5.0,
-      "eval_runtime": 0.8101,
-      "eval_samples_per_second": 164.178,
-      "eval_steps_per_second": 20.985,
+      "eval_accuracy": 0.3308270676691729,
+      "eval_loss": 1.1081749200820923,
+      "eval_runtime": 0.9355,
+      "eval_samples_per_second": 142.176,
+      "eval_steps_per_second": 18.173,
       "step": 650
     },
     {
       "epoch": 5.076923076923077,
-      "grad_norm": 0.30358830094337463,
+      "grad_norm": 0.6719330549240112,
       "learning_rate": 0.0024615384615384616,
-      "loss": 1.1044,
+      "loss": 1.1054,
       "step": 660
     },
     {
       "epoch": 5.153846153846154,
-      "grad_norm": 0.42961689829826355,
+      "grad_norm": 0.8374503254890442,
       "learning_rate": 0.002423076923076923,
-      "loss": 1.0977,
+      "loss": 1.1134,
       "step": 670
     },
     {
       "epoch": 5.230769230769231,
-      "grad_norm": 0.24340471625328064,
+      "grad_norm": 0.5713789463043213,
       "learning_rate": 0.0023846153846153848,
-      "loss": 1.1379,
+      "loss": 1.1534,
       "step": 680
     },
     {
       "epoch": 5.3076923076923075,
-      "grad_norm": 0.9648124575614929,
+      "grad_norm": 1.0185028314590454,
       "learning_rate": 0.0023461538461538463,
-      "loss": 1.1314,
+      "loss": 1.1368,
       "step": 690
     },
     {
       "epoch": 5.384615384615385,
-      "grad_norm": 0.6379570960998535,
+      "grad_norm": 1.021331787109375,
       "learning_rate": 0.002307692307692308,
-      "loss": 1.1104,
+      "loss": 1.094,
       "step": 700
     },
     {
       "epoch": 5.461538461538462,
-      "grad_norm": 0.32133665680885315,
+      "grad_norm": 0.9150602221488953,
       "learning_rate": 0.0022692307692307695,
-      "loss": 1.1061,
+      "loss": 1.1187,
       "step": 710
     },
     {
       "epoch": 5.538461538461538,
-      "grad_norm": 0.314644455909729,
+      "grad_norm": 0.741632878780365,
       "learning_rate": 0.002230769230769231,
-      "loss": 1.1035,
+      "loss": 1.1129,
       "step": 720
     },
     {
       "epoch": 5.615384615384615,
-      "grad_norm": 0.4097970128059387,
+      "grad_norm": 1.5859191417694092,
       "learning_rate": 0.002192307692307692,
-      "loss": 1.1035,
+      "loss": 1.135,
       "step": 730
     },
     {
       "epoch": 5.6923076923076925,
-      "grad_norm": 0.3651571273803711,
+      "grad_norm": 2.2667315006256104,
       "learning_rate": 0.002153846153846154,
-      "loss": 1.1071,
+      "loss": 1.1474,
       "step": 740
     },
     {
       "epoch": 5.769230769230769,
-      "grad_norm": 0.30447086691856384,
+      "grad_norm": 1.3794482946395874,
       "learning_rate": 0.0021153846153846153,
-      "loss": 1.1048,
+      "loss": 1.1095,
       "step": 750
     },
     {
       "epoch": 5.846153846153846,
-      "grad_norm": 0.35882461071014404,
+      "grad_norm": 1.2423471212387085,
       "learning_rate": 0.002076923076923077,
-      "loss": 1.106,
+      "loss": 1.1057,
       "step": 760
     },
     {
       "epoch": 5.923076923076923,
-      "grad_norm": 0.6777219772338867,
+      "grad_norm": 4.28483247756958,
       "learning_rate": 0.0020384615384615385,
-      "loss": 1.101,
+      "loss": 1.1093,
       "step": 770
     },
     {
       "epoch": 6.0,
-      "grad_norm": 1.031847357749939,
+      "grad_norm": 4.998449802398682,
       "learning_rate": 0.002,
-      "loss": 1.1027,
+      "loss": 1.1748,
       "step": 780
     },
     {
       "epoch": 6.0,
-      "eval_runtime": 0.7636,
-      "eval_samples_per_second": 174.172,
-      "eval_steps_per_second": 22.263,
+      "eval_accuracy": 0.3308270676691729,
+      "eval_loss": 1.1041723489761353,
+      "eval_runtime": 0.9442,
+      "eval_samples_per_second": 140.864,
+      "eval_steps_per_second": 18.005,
       "step": 780
     },
     {
       "epoch": 6.076923076923077,
-      "grad_norm": 0.43265023827552795,
+      "grad_norm": 2.81730318069458,
       "learning_rate": 0.0019615384615384616,
-      "loss": 1.104,
+      "loss": 1.1381,
       "step": 790
     },
     {
       "epoch": 6.153846153846154,
-      "grad_norm": 0.4353933334350586,
+      "grad_norm": 2.5270557403564453,
       "learning_rate": 0.0019230769230769232,
-      "loss": 1.099,
+      "loss": 1.1104,
       "step": 800
     },
     {
       "epoch": 6.230769230769231,
-      "grad_norm": 0.11455998569726944,
+      "grad_norm": 0.8537469506263733,
       "learning_rate": 0.0018846153846153845,
-      "loss": 1.0985,
+      "loss": 1.1327,
       "step": 810
     },
     {
       "epoch": 6.3076923076923075,
-      "grad_norm": 0.1155104711651802,
+      "grad_norm": 0.8818097710609436,
       "learning_rate": 0.0018461538461538463,
-      "loss": 1.1026,
+      "loss": 1.1437,
       "step": 820
     },
     {
       "epoch": 6.384615384615385,
-      "grad_norm": 0.44076037406921387,
+      "grad_norm": 2.496903896331787,
       "learning_rate": 0.0018076923076923077,
-      "loss": 1.0985,
+      "loss": 1.1251,
       "step": 830
     },
     {
       "epoch": 6.461538461538462,
-      "grad_norm": 0.5386171936988831,
+      "grad_norm": 3.2999167442321777,
       "learning_rate": 0.0017692307692307695,
-      "loss": 1.1025,
+      "loss": 1.1373,
       "step": 840
     },
     {
       "epoch": 6.538461538461538,
-      "grad_norm": 0.611710786819458,
+      "grad_norm": 2.9537723064422607,
       "learning_rate": 0.0017307692307692308,
-      "loss": 1.0992,
+      "loss": 1.176,
       "step": 850
     },
     {
       "epoch": 6.615384615384615,
-      "grad_norm": 0.2715957760810852,
+      "grad_norm": 1.5148764848709106,
       "learning_rate": 0.0016923076923076924,
-      "loss": 1.1008,
+      "loss": 1.1146,
       "step": 860
     },
     {
       "epoch": 6.6923076923076925,
-      "grad_norm": 0.4223721921443939,
+      "grad_norm": 1.3614485263824463,
       "learning_rate": 0.0016538461538461537,
-      "loss": 1.089,
+      "loss": 1.0895,
       "step": 870
     },
     {
       "epoch": 6.769230769230769,
-      "grad_norm": 0.14638401567935944,
+      "grad_norm": 0.6686407327651978,
       "learning_rate": 0.0016153846153846155,
-      "loss": 1.1237,
+      "loss": 1.1245,
       "step": 880
     },
     {
       "epoch": 6.846153846153846,
-      "grad_norm": 0.32549309730529785,
+      "grad_norm": 1.6276365518569946,
       "learning_rate": 0.0015769230769230769,
-      "loss": 1.1031,
+      "loss": 1.1412,
       "step": 890
     },
     {
       "epoch": 6.923076923076923,
-      "grad_norm": 0.4604528844356537,
+      "grad_norm": 1.4945542812347412,
       "learning_rate": 0.0015384615384615387,
-      "loss": 1.1001,
+      "loss": 1.1247,
       "step": 900
     },
     {
       "epoch": 7.0,
-      "grad_norm": 0.5566179156303406,
+      "grad_norm": 2.74285888671875,
       "learning_rate": 0.0015,
-      "loss": 1.0954,
+      "loss": 1.1514,
       "step": 910
     },
     {
       "epoch": 7.0,
-      "eval_runtime": 0.7734,
-      "eval_samples_per_second": 171.968,
-      "eval_steps_per_second": 21.981,
+      "eval_accuracy": 0.3308270676691729,
+      "eval_loss": 1.1007708311080933,
+      "eval_runtime": 0.9264,
+      "eval_samples_per_second": 143.562,
+      "eval_steps_per_second": 18.35,
       "step": 910
     },
     {
       "epoch": 7.076923076923077,
-      "grad_norm": 0.5109624266624451,
+      "grad_norm": 2.248867988586426,
       "learning_rate": 0.0014615384615384616,
-      "loss": 1.1043,
+      "loss": 1.0913,
       "step": 920
     },
     {
       "epoch": 7.153846153846154,
-      "grad_norm": 0.5437686443328857,
+      "grad_norm": 2.5912275314331055,
       "learning_rate": 0.001423076923076923,
-      "loss": 1.1055,
+      "loss": 1.1193,
       "step": 930
     },
     {
       "epoch": 7.230769230769231,
-      "grad_norm": 0.6647012233734131,
+      "grad_norm": 3.573967695236206,
       "learning_rate": 0.0013846153846153847,
-      "loss": 1.1021,
+      "loss": 1.1352,
       "step": 940
     },
     {
       "epoch": 7.3076923076923075,
-      "grad_norm": 0.4888548254966736,
+      "grad_norm": 2.077554225921631,
       "learning_rate": 0.001346153846153846,
-      "loss": 1.0968,
+      "loss": 1.0947,
       "step": 950
     },
     {
       "epoch": 7.384615384615385,
-      "grad_norm": 0.26103150844573975,
+      "grad_norm": 1.6364086866378784,
       "learning_rate": 0.0013076923076923079,
-      "loss": 1.104,
+      "loss": 1.1296,
       "step": 960
     },
     {
       "epoch": 7.461538461538462,
-      "grad_norm": 0.2501744031906128,
+      "grad_norm": 1.1806069612503052,
       "learning_rate": 0.0012692307692307692,
-      "loss": 1.1008,
+      "loss": 1.1033,
       "step": 970
     },
     {
       "epoch": 7.538461538461538,
-      "grad_norm": 0.37639835476875305,
+      "grad_norm": 2.4788713455200195,
       "learning_rate": 0.0012307692307692308,
-      "loss": 1.0998,
+      "loss": 1.1148,
       "step": 980
     },
     {
       "epoch": 7.615384615384615,
-      "grad_norm": 0.23087725043296814,
+      "grad_norm": 0.8245519399642944,
       "learning_rate": 0.0011923076923076924,
-      "loss": 1.0979,
+      "loss": 1.1121,
       "step": 990
     },
     {
       "epoch": 7.6923076923076925,
-      "grad_norm": 0.3334718644618988,
+      "grad_norm": 1.568903923034668,
       "learning_rate": 0.001153846153846154,
-      "loss": 1.1068,
+      "loss": 1.1163,
       "step": 1000
     },
     {
       "epoch": 7.769230769230769,
-      "grad_norm": 0.5154008865356445,
+      "grad_norm": 2.2075676918029785,
       "learning_rate": 0.0011153846153846155,
-      "loss": 1.099,
+      "loss": 1.0956,
       "step": 1010
     },
     {
       "epoch": 7.846153846153846,
-      "grad_norm": 0.2505805492401123,
+      "grad_norm": 1.0866578817367554,
       "learning_rate": 0.001076923076923077,
-      "loss": 1.0985,
+      "loss": 1.1065,
       "step": 1020
     },
     {
       "epoch": 7.923076923076923,
-      "grad_norm": 0.47592583298683167,
+      "grad_norm": 2.4340384006500244,
       "learning_rate": 0.0010384615384615384,
-      "loss": 1.0968,
+      "loss": 1.1118,
       "step": 1030
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.4868021309375763,
+      "grad_norm": 2.1427950859069824,
       "learning_rate": 0.001,
-      "loss": 1.1047,
+      "loss": 1.0956,
       "step": 1040
     },
     {
       "epoch": 8.0,
-      "eval_runtime": 0.7758,
-      "eval_samples_per_second": 171.434,
-      "eval_steps_per_second": 21.913,
+      "eval_accuracy": 0.3383458646616541,
+      "eval_loss": 1.1004995107650757,
+      "eval_runtime": 0.9133,
+      "eval_samples_per_second": 145.621,
+      "eval_steps_per_second": 18.613,
       "step": 1040
     },
     {
       "epoch": 8.076923076923077,
-      "grad_norm": 0.12491131573915482,
+      "grad_norm": 0.689738392829895,
       "learning_rate": 0.0009615384615384616,
-      "loss": 1.0981,
+      "loss": 1.1119,
       "step": 1050
     },
     {
       "epoch": 8.153846153846153,
-      "grad_norm": 0.23577113449573517,
+      "grad_norm": 0.7051054835319519,
       "learning_rate": 0.0009230769230769232,
-      "loss": 1.0998,
+      "loss": 1.1076,
       "step": 1060
     },
     {
       "epoch": 8.23076923076923,
-      "grad_norm": 0.12919014692306519,
+      "grad_norm": 0.8233575820922852,
       "learning_rate": 0.0008846153846153847,
-      "loss": 1.098,
+      "loss": 1.1,
       "step": 1070
     },
     {
       "epoch": 8.307692307692308,
-      "grad_norm": 0.11554103344678879,
+      "grad_norm": 0.4401038885116577,
       "learning_rate": 0.0008461538461538462,
-      "loss": 1.099,
+      "loss": 1.1231,
       "step": 1080
     },
     {
       "epoch": 8.384615384615385,
-      "grad_norm": 0.3568715751171112,
+      "grad_norm": 2.0618515014648438,
       "learning_rate": 0.0008076923076923078,
-      "loss": 1.0935,
+      "loss": 1.0877,
       "step": 1090
     },
     {
       "epoch": 8.461538461538462,
-      "grad_norm": 0.4273003041744232,
+      "grad_norm": 1.629764199256897,
       "learning_rate": 0.0007692307692307693,
-      "loss": 1.1025,
+      "loss": 1.1165,
       "step": 1100
     },
     {
       "epoch": 8.538461538461538,
-      "grad_norm": 0.11346741020679474,
+      "grad_norm": 0.9111972451210022,
       "learning_rate": 0.0007307692307692308,
-      "loss": 1.1079,
+      "loss": 1.1033,
       "step": 1110
     },
     {
       "epoch": 8.615384615384615,
-      "grad_norm": 0.5924614667892456,
+      "grad_norm": 2.03430438041687,
       "learning_rate": 0.0006923076923076924,
-      "loss": 1.095,
+      "loss": 1.0906,
       "step": 1120
     },
     {
       "epoch": 8.692307692307692,
-      "grad_norm": 0.8924316763877869,
+      "grad_norm": 3.2121763229370117,
       "learning_rate": 0.0006538461538461539,
-      "loss": 1.0969,
+      "loss": 1.126,
       "step": 1130
     },
     {
       "epoch": 8.76923076923077,
-      "grad_norm": 0.14277611672878265,
+      "grad_norm": 0.34755992889404297,
       "learning_rate": 0.0006153846153846154,
-      "loss": 1.112,
+      "loss": 1.1051,
       "step": 1140
     },
     {
       "epoch": 8.846153846153847,
-      "grad_norm": 0.29049235582351685,
+      "grad_norm": 0.7230011224746704,
       "learning_rate": 0.000576923076923077,
-      "loss": 1.0914,
+      "loss": 1.0721,
       "step": 1150
     },
     {
       "epoch": 8.923076923076923,
-      "grad_norm": 0.5114134550094604,
+      "grad_norm": 2.8672590255737305,
       "learning_rate": 0.0005384615384615385,
-      "loss": 1.1061,
+      "loss": 1.0711,
       "step": 1160
     },
     {
       "epoch": 9.0,
-      "grad_norm": 1.142321228981018,
+      "grad_norm": 2.920103073120117,
       "learning_rate": 0.0005,
-      "loss": 1.1105,
+      "loss": 1.1448,
       "step": 1170
     },
     {
       "epoch": 9.0,
-      "eval_runtime": 0.7827,
-      "eval_samples_per_second": 169.933,
-      "eval_steps_per_second": 21.721,
+      "eval_accuracy": 0.3308270676691729,
+      "eval_loss": 1.111855149269104,
+      "eval_runtime": 0.9175,
+      "eval_samples_per_second": 144.96,
+      "eval_steps_per_second": 18.529,
       "step": 1170
     },
     {
       "epoch": 9.076923076923077,
-      "grad_norm": 0.12881995737552643,
+      "grad_norm": 0.26905354857444763,
       "learning_rate": 0.0004615384615384616,
-      "loss": 1.0994,
+      "loss": 1.1121,
       "step": 1180
     },
     {
       "epoch": 9.153846153846153,
-      "grad_norm": 0.12638165056705475,
+      "grad_norm": 0.22876305878162384,
       "learning_rate": 0.0004230769230769231,
-      "loss": 1.0982,
+      "loss": 1.0967,
       "step": 1190
     },
     {
       "epoch": 9.23076923076923,
-      "grad_norm": 0.13503584265708923,
+      "grad_norm": 0.24429693818092346,
       "learning_rate": 0.00038461538461538467,
-      "loss": 1.0996,
+      "loss": 1.1023,
       "step": 1200
     },
     {
       "epoch": 9.307692307692308,
-      "grad_norm": 0.23928503692150116,
+      "grad_norm": 0.6317471861839294,
       "learning_rate": 0.0003461538461538462,
-      "loss": 1.1,
+      "loss": 1.1062,
       "step": 1210
     },
     {
       "epoch": 9.384615384615385,
-      "grad_norm": 0.14712156355381012,
+      "grad_norm": 0.41996484994888306,
       "learning_rate": 0.0003076923076923077,
-      "loss": 1.1033,
+      "loss": 1.1162,
       "step": 1220
     },
     {
       "epoch": 9.461538461538462,
-      "grad_norm": 0.603887140750885,
+      "grad_norm": 1.6321758031845093,
       "learning_rate": 0.0002692307692307693,
-      "loss": 1.0979,
+      "loss": 1.0976,
       "step": 1230
     },
     {
       "epoch": 9.538461538461538,
-      "grad_norm": 0.34505996108055115,
+      "grad_norm": 0.966174840927124,
       "learning_rate": 0.0002307692307692308,
-      "loss": 1.0977,
+      "loss": 1.0979,
       "step": 1240
     },
     {
       "epoch": 9.615384615384615,
-      "grad_norm": 0.11722344160079956,
+      "grad_norm": 0.23766568303108215,
       "learning_rate": 0.00019230769230769233,
-      "loss": 1.1001,
+      "loss": 1.1012,
       "step": 1250
     },
     {
       "epoch": 9.692307692307692,
-      "grad_norm": 0.33938324451446533,
+      "grad_norm": 0.930920422077179,
       "learning_rate": 0.00015384615384615385,
-      "loss": 1.0988,
+      "loss": 1.0991,
       "step": 1260
     },
     {
       "epoch": 9.76923076923077,
-      "grad_norm": 0.1380094438791275,
+      "grad_norm": 0.34004735946655273,
       "learning_rate": 0.0001153846153846154,
-      "loss": 1.1004,
+      "loss": 1.1009,
       "step": 1270
     },
     {
       "epoch": 9.846153846153847,
-      "grad_norm": 0.2527919411659241,
+      "grad_norm": 0.663216233253479,
       "learning_rate": 7.692307692307693e-05,
-      "loss": 1.0999,
+      "loss": 1.0997,
       "step": 1280
     },
     {
       "epoch": 9.923076923076923,
-      "grad_norm": 0.316256582736969,
+      "grad_norm": 0.9082524180412292,
       "learning_rate": 3.846153846153846e-05,
-      "loss": 1.0984,
+      "loss": 1.0998,
       "step": 1290
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.4816994071006775,
+      "grad_norm": 1.356846570968628,
       "learning_rate": 0.0,
-      "loss": 1.0979,
+      "loss": 1.0991,
       "step": 1300
     },
     {
       "epoch": 10.0,
-      "eval_runtime": 0.7803,
-      "eval_samples_per_second": 170.446,
-      "eval_steps_per_second": 21.786,
+      "eval_accuracy": 0.3308270676691729,
+      "eval_loss": 1.0983037948608398,
+      "eval_runtime": 0.9323,
+      "eval_samples_per_second": 142.651,
+      "eval_steps_per_second": 18.234,
       "step": 1300
     },
     {
       "epoch": 10.0,
       "step": 1300,
       "total_flos": 8.400578669044531e+17,
-      "train_loss": 0.9780362775692573,
-      "train_runtime": 137.5902,
-      "train_samples_per_second": 75.151,
-      "train_steps_per_second": 9.448
+      "train_loss": 0.9618942783429072,
+      "train_runtime": 164.295,
+      "train_samples_per_second": 62.936,
+      "train_steps_per_second": 7.913
     }
   ],
   "logging_steps": 10,