🍻 cheers

README.md

@@ -2,6 +2,7 @@
 license: apache-2.0
 base_model: google/vit-base-patch16-224-in21k
 tags:
+- image-classification
 - generated_from_trainer
 datasets:
 - renovation
@@ -14,7 +15,7 @@ model-index:
       name: Image Classification
       type: image-classification
     dataset:
-      name: renovation
+      name: renovations
       type: renovation
       config: default
       split: validation
@@ -22,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.7031963470319634
+      value: 0.6666666666666666
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +31,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vit-base-renovation2
 
-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 renovation 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 renovations dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.9788
-- Accuracy: 0.7032
+- Loss: 0.8273
+- Accuracy: 0.6667
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.6027397260273972,
-    "eval_loss": 1.2383606433868408,
-    "eval_runtime": 7.929,
-    "eval_samples_per_second": 27.62,
-    "eval_steps_per_second": 3.531,
+    "eval_accuracy": 0.6666666666666666,
+    "eval_loss": 0.8272958397865295,
+    "eval_runtime": 7.1557,
+    "eval_samples_per_second": 30.605,
+    "eval_steps_per_second": 3.913,
     "total_flos": 6.10974224738132e+17,
-    "train_loss": 0.14545002643529686,
-    "train_runtime": 906.411,
-    "train_samples_per_second": 8.698,
-    "train_steps_per_second": 0.547
+    "train_loss": 0.601069396421794,
+    "train_runtime": 693.9158,
+    "train_samples_per_second": 11.362,
+    "train_steps_per_second": 0.715
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.6027397260273972,
-    "eval_loss": 1.2383606433868408,
-    "eval_runtime": 7.929,
-    "eval_samples_per_second": 27.62,
-    "eval_steps_per_second": 3.531
+    "eval_accuracy": 0.6666666666666666,
+    "eval_loss": 0.8272958397865295,
+    "eval_runtime": 7.1557,
+    "eval_samples_per_second": 30.605,
+    "eval_steps_per_second": 3.913
 }

runs/Mar22_21-16-36_9f0b864d5439/events.out.tfevents.1711142925.9f0b864d5439.318.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a41f191e74b8229e363f237dafd5ee6c71efac8ffb5ab1cc41c09f5f0a5bfed6
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
     "total_flos": 6.10974224738132e+17,
-    "train_loss": 0.14545002643529686,
-    "train_runtime": 906.411,
-    "train_samples_per_second": 8.698,
-    "train_steps_per_second": 0.547
+    "train_loss": 0.601069396421794,
+    "train_runtime": 693.9158,
+    "train_samples_per_second": 11.362,
+    "train_steps_per_second": 0.715
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 1.2383606433868408,
-  "best_model_checkpoint": "./vit-base-renovation2/checkpoint-25",
+  "best_metric": 0.8272958397865295,
+  "best_model_checkpoint": "./vit-base-renovation2/checkpoint-250",
   "epoch": 4.0,
   "eval_steps": 25,
   "global_step": 496,
@@ -10,526 +10,526 @@
   "log_history": [
     {
       "epoch": 0.08,
-      "grad_norm": 2.575977325439453,
+      "grad_norm": 1.815894365310669,
       "learning_rate": 0.00019596774193548388,
-      "loss": 0.3595,
+      "loss": 1.5846,
       "step": 10
     },
     {
       "epoch": 0.16,
-      "grad_norm": 0.9806227684020996,
+      "grad_norm": 2.0016415119171143,
       "learning_rate": 0.00019193548387096775,
-      "loss": 0.273,
+      "loss": 1.359,
       "step": 20
     },
     {
       "epoch": 0.2,
-      "eval_accuracy": 0.6027397260273972,
-      "eval_loss": 1.2383606433868408,
-      "eval_runtime": 7.9057,
-      "eval_samples_per_second": 27.702,
-      "eval_steps_per_second": 3.542,
+      "eval_accuracy": 0.4657534246575342,
+      "eval_loss": 1.2074170112609863,
+      "eval_runtime": 7.6692,
+      "eval_samples_per_second": 28.556,
+      "eval_steps_per_second": 3.651,
       "step": 25
     },
     {
       "epoch": 0.24,
-      "grad_norm": 2.519897699356079,
+      "grad_norm": 1.6700384616851807,
       "learning_rate": 0.00018790322580645164,
-      "loss": 0.3335,
+      "loss": 1.182,
       "step": 30
     },
     {
       "epoch": 0.32,
-      "grad_norm": 4.2508063316345215,
+      "grad_norm": 2.4473767280578613,
       "learning_rate": 0.00018387096774193548,
-      "loss": 0.3046,
+      "loss": 1.1559,
       "step": 40
     },
     {
       "epoch": 0.4,
-      "grad_norm": 1.2349441051483154,
+      "grad_norm": 2.757627010345459,
       "learning_rate": 0.00017983870967741935,
-      "loss": 0.5153,
+      "loss": 1.1384,
       "step": 50
     },
     {
       "epoch": 0.4,
-      "eval_accuracy": 0.5844748858447488,
-      "eval_loss": 1.40598464012146,
-      "eval_runtime": 8.1033,
-      "eval_samples_per_second": 27.026,
-      "eval_steps_per_second": 3.455,
+      "eval_accuracy": 0.5205479452054794,
+      "eval_loss": 1.1213266849517822,
+      "eval_runtime": 8.2359,
+      "eval_samples_per_second": 26.591,
+      "eval_steps_per_second": 3.4,
       "step": 50
     },
     {
       "epoch": 0.48,
-      "grad_norm": 2.6730687618255615,
+      "grad_norm": 1.8923784494400024,
       "learning_rate": 0.00017580645161290325,
-      "loss": 0.3137,
+      "loss": 1.147,
       "step": 60
     },
     {
       "epoch": 0.56,
-      "grad_norm": 4.323659420013428,
+      "grad_norm": 2.0101828575134277,
       "learning_rate": 0.00017177419354838711,
-      "loss": 0.2792,
+      "loss": 1.0866,
       "step": 70
     },
     {
       "epoch": 0.6,
-      "eval_accuracy": 0.593607305936073,
-      "eval_loss": 1.30255126953125,
-      "eval_runtime": 7.4143,
-      "eval_samples_per_second": 29.538,
-      "eval_steps_per_second": 3.776,
+      "eval_accuracy": 0.6301369863013698,
+      "eval_loss": 0.9746416807174683,
+      "eval_runtime": 8.0785,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 3.466,
       "step": 75
     },
     {
       "epoch": 0.65,
-      "grad_norm": 1.2822186946868896,
+      "grad_norm": 3.479001522064209,
       "learning_rate": 0.00016774193548387098,
-      "loss": 0.3751,
+      "loss": 0.875,
       "step": 80
     },
     {
       "epoch": 0.73,
-      "grad_norm": 3.15390944480896,
+      "grad_norm": 3.4011082649230957,
       "learning_rate": 0.00016370967741935485,
-      "loss": 0.5094,
+      "loss": 1.1928,
       "step": 90
     },
     {
       "epoch": 0.81,
-      "grad_norm": 3.2424371242523193,
-      "learning_rate": 0.00016008064516129033,
-      "loss": 0.5516,
+      "grad_norm": 2.3756794929504395,
+      "learning_rate": 0.00015967741935483872,
+      "loss": 1.1787,
       "step": 100
     },
     {
       "epoch": 0.81,
-      "eval_accuracy": 0.6027397260273972,
-      "eval_loss": 1.3998719453811646,
-      "eval_runtime": 7.9973,
-      "eval_samples_per_second": 27.384,
-      "eval_steps_per_second": 3.501,
+      "eval_accuracy": 0.5662100456621004,
+      "eval_loss": 1.0523391962051392,
+      "eval_runtime": 9.1559,
+      "eval_samples_per_second": 23.919,
+      "eval_steps_per_second": 3.058,
       "step": 100
     },
     {
       "epoch": 0.89,
-      "grad_norm": 1.8185967206954956,
-      "learning_rate": 0.0001560483870967742,
-      "loss": 0.2684,
+      "grad_norm": 1.5179095268249512,
+      "learning_rate": 0.0001556451612903226,
+      "loss": 0.9827,
       "step": 110
     },
     {
       "epoch": 0.97,
-      "grad_norm": 3.6117029190063477,
-      "learning_rate": 0.0001520161290322581,
-      "loss": 0.4247,
+      "grad_norm": 2.3550424575805664,
+      "learning_rate": 0.00015161290322580646,
+      "loss": 0.9242,
       "step": 120
     },
     {
       "epoch": 1.01,
-      "eval_accuracy": 0.5981735159817352,
-      "eval_loss": 1.2620913982391357,
-      "eval_runtime": 8.1833,
-      "eval_samples_per_second": 26.762,
-      "eval_steps_per_second": 3.422,
+      "eval_accuracy": 0.6255707762557078,
+      "eval_loss": 0.9543189406394958,
+      "eval_runtime": 7.3462,
+      "eval_samples_per_second": 29.812,
+      "eval_steps_per_second": 3.812,
       "step": 125
     },
     {
       "epoch": 1.05,
-      "grad_norm": 0.7001488208770752,
-      "learning_rate": 0.00014798387096774193,
-      "loss": 0.1693,
+      "grad_norm": 1.8633378744125366,
+      "learning_rate": 0.00014758064516129032,
+      "loss": 0.8227,
       "step": 130
     },
     {
       "epoch": 1.13,
-      "grad_norm": 3.214643716812134,
-      "learning_rate": 0.0001439516129032258,
-      "loss": 0.2085,
+      "grad_norm": 2.0424883365631104,
+      "learning_rate": 0.00014354838709677422,
+      "loss": 0.747,
       "step": 140
     },
     {
       "epoch": 1.21,
-      "grad_norm": 2.9545109272003174,
-      "learning_rate": 0.00013991935483870967,
-      "loss": 0.1556,
+      "grad_norm": 2.2865993976593018,
+      "learning_rate": 0.0001395161290322581,
+      "loss": 0.7945,
       "step": 150
     },
     {
       "epoch": 1.21,
-      "eval_accuracy": 0.5570776255707762,
-      "eval_loss": 1.5660632848739624,
-      "eval_runtime": 8.1924,
-      "eval_samples_per_second": 26.732,
-      "eval_steps_per_second": 3.418,
+      "eval_accuracy": 0.6118721461187214,
+      "eval_loss": 0.9199973940849304,
+      "eval_runtime": 7.7884,
+      "eval_samples_per_second": 28.119,
+      "eval_steps_per_second": 3.595,
       "step": 150
     },
     {
       "epoch": 1.29,
-      "grad_norm": 3.702380418777466,
-      "learning_rate": 0.00013588709677419357,
-      "loss": 0.1262,
+      "grad_norm": 1.884401798248291,
+      "learning_rate": 0.00013548387096774193,
+      "loss": 0.9408,
       "step": 160
     },
     {
       "epoch": 1.37,
-      "grad_norm": 2.634195327758789,
-      "learning_rate": 0.00013185483870967743,
-      "loss": 0.1458,
+      "grad_norm": 2.085380792617798,
+      "learning_rate": 0.0001314516129032258,
+      "loss": 0.8379,
       "step": 170
     },
     {
       "epoch": 1.41,
-      "eval_accuracy": 0.634703196347032,
-      "eval_loss": 1.3458657264709473,
-      "eval_runtime": 7.2989,
-      "eval_samples_per_second": 30.005,
-      "eval_steps_per_second": 3.836,
+      "eval_accuracy": 0.6712328767123288,
+      "eval_loss": 0.8447285294532776,
+      "eval_runtime": 7.8794,
+      "eval_samples_per_second": 27.794,
+      "eval_steps_per_second": 3.554,
       "step": 175
     },
     {
       "epoch": 1.45,
-      "grad_norm": 1.4022108316421509,
-      "learning_rate": 0.0001278225806451613,
-      "loss": 0.0859,
+      "grad_norm": 2.1437325477600098,
+      "learning_rate": 0.0001274193548387097,
+      "loss": 0.8348,
       "step": 180
     },
     {
       "epoch": 1.53,
-      "grad_norm": 6.311899185180664,
-      "learning_rate": 0.00012379032258064514,
-      "loss": 0.0868,
+      "grad_norm": 1.6616462469100952,
+      "learning_rate": 0.00012338709677419356,
+      "loss": 0.6521,
       "step": 190
     },
     {
       "epoch": 1.61,
-      "grad_norm": 1.674466609954834,
-      "learning_rate": 0.00011975806451612903,
-      "loss": 0.1595,
+      "grad_norm": 2.3329973220825195,
+      "learning_rate": 0.00011935483870967743,
+      "loss": 0.7253,
       "step": 200
     },
     {
       "epoch": 1.61,
-      "eval_accuracy": 0.5981735159817352,
-      "eval_loss": 1.5277855396270752,
-      "eval_runtime": 8.087,
-      "eval_samples_per_second": 27.081,
-      "eval_steps_per_second": 3.462,
+      "eval_accuracy": 0.6575342465753424,
+      "eval_loss": 0.8642416596412659,
+      "eval_runtime": 7.2392,
+      "eval_samples_per_second": 30.252,
+      "eval_steps_per_second": 3.868,
       "step": 200
     },
     {
       "epoch": 1.69,
-      "grad_norm": 1.2903810739517212,
-      "learning_rate": 0.00011572580645161291,
-      "loss": 0.1657,
+      "grad_norm": 3.4008474349975586,
+      "learning_rate": 0.00011532258064516131,
+      "loss": 0.6683,
       "step": 210
     },
     {
       "epoch": 1.77,
-      "grad_norm": 1.2086151838302612,
-      "learning_rate": 0.00011169354838709678,
-      "loss": 0.1195,
+      "grad_norm": 2.7154157161712646,
+      "learning_rate": 0.00011129032258064515,
+      "loss": 0.6344,
       "step": 220
     },
     {
       "epoch": 1.81,
-      "eval_accuracy": 0.6255707762557078,
-      "eval_loss": 1.5303133726119995,
-      "eval_runtime": 7.5329,
-      "eval_samples_per_second": 29.072,
-      "eval_steps_per_second": 3.717,
+      "eval_accuracy": 0.6438356164383562,
+      "eval_loss": 0.8443467020988464,
+      "eval_runtime": 7.7267,
+      "eval_samples_per_second": 28.343,
+      "eval_steps_per_second": 3.624,
       "step": 225
     },
     {
       "epoch": 1.85,
-      "grad_norm": 7.646343231201172,
-      "learning_rate": 0.00010766129032258066,
-      "loss": 0.1381,
+      "grad_norm": 2.2653236389160156,
+      "learning_rate": 0.00010725806451612903,
+      "loss": 0.768,
       "step": 230
     },
     {
       "epoch": 1.94,
-      "grad_norm": 5.495790481567383,
-      "learning_rate": 0.00010362903225806453,
-      "loss": 0.1761,
+      "grad_norm": 2.3153140544891357,
+      "learning_rate": 0.0001032258064516129,
+      "loss": 0.8194,
       "step": 240
     },
     {
       "epoch": 2.02,
-      "grad_norm": 0.2532494366168976,
-      "learning_rate": 9.95967741935484e-05,
-      "loss": 0.1507,
+      "grad_norm": 1.3728018999099731,
+      "learning_rate": 9.919354838709678e-05,
+      "loss": 0.6521,
       "step": 250
     },
     {
       "epoch": 2.02,
-      "eval_accuracy": 0.5844748858447488,
-      "eval_loss": 1.7701371908187866,
-      "eval_runtime": 8.229,
-      "eval_samples_per_second": 26.613,
-      "eval_steps_per_second": 3.403,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 0.8272958397865295,
+      "eval_runtime": 7.9935,
+      "eval_samples_per_second": 27.397,
+      "eval_steps_per_second": 3.503,
       "step": 250
     },
     {
       "epoch": 2.1,
-      "grad_norm": 2.001723527908325,
-      "learning_rate": 9.556451612903226e-05,
-      "loss": 0.141,
+      "grad_norm": 1.0636628866195679,
+      "learning_rate": 9.516129032258065e-05,
+      "loss": 0.3979,
       "step": 260
     },
     {
       "epoch": 2.18,
-      "grad_norm": 0.07203235477209091,
-      "learning_rate": 9.153225806451613e-05,
-      "loss": 0.023,
+      "grad_norm": 1.3106963634490967,
+      "learning_rate": 9.112903225806452e-05,
+      "loss": 0.3627,
       "step": 270
     },
     {
       "epoch": 2.22,
-      "eval_accuracy": 0.6301369863013698,
-      "eval_loss": 1.535449743270874,
-      "eval_runtime": 7.2739,
-      "eval_samples_per_second": 30.108,
-      "eval_steps_per_second": 3.849,
+      "eval_accuracy": 0.6712328767123288,
+      "eval_loss": 0.865295946598053,
+      "eval_runtime": 7.9627,
+      "eval_samples_per_second": 27.503,
+      "eval_steps_per_second": 3.516,
       "step": 275
     },
     {
       "epoch": 2.26,
-      "grad_norm": 0.10443231463432312,
-      "learning_rate": 8.75e-05,
-      "loss": 0.0434,
+      "grad_norm": 1.7618376016616821,
+      "learning_rate": 8.709677419354839e-05,
+      "loss": 0.3624,
       "step": 280
     },
     {
       "epoch": 2.34,
-      "grad_norm": 1.9513520002365112,
-      "learning_rate": 8.346774193548388e-05,
-      "loss": 0.0221,
+      "grad_norm": 1.3475382328033447,
+      "learning_rate": 8.306451612903227e-05,
+      "loss": 0.4034,
       "step": 290
     },
     {
       "epoch": 2.42,
-      "grad_norm": 4.193704605102539,
-      "learning_rate": 7.943548387096774e-05,
-      "loss": 0.028,
+      "grad_norm": 3.0443389415740967,
+      "learning_rate": 7.903225806451613e-05,
+      "loss": 0.2523,
       "step": 300
     },
     {
       "epoch": 2.42,
-      "eval_accuracy": 0.6301369863013698,
-      "eval_loss": 1.6535084247589111,
-      "eval_runtime": 8.3156,
-      "eval_samples_per_second": 26.336,
-      "eval_steps_per_second": 3.367,
+      "eval_accuracy": 0.6894977168949772,
+      "eval_loss": 0.874794065952301,
+      "eval_runtime": 7.2976,
+      "eval_samples_per_second": 30.01,
+      "eval_steps_per_second": 3.837,
       "step": 300
     },
     {
       "epoch": 2.5,
-      "grad_norm": 0.11177629232406616,
-      "learning_rate": 7.540322580645162e-05,
-      "loss": 0.0573,
+      "grad_norm": 2.961167812347412,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.3129,
       "step": 310
     },
     {
       "epoch": 2.58,
-      "grad_norm": 0.0510709173977375,
-      "learning_rate": 7.137096774193549e-05,
-      "loss": 0.0698,
+      "grad_norm": 1.5659066438674927,
+      "learning_rate": 7.096774193548388e-05,
+      "loss": 0.363,
       "step": 320
     },
     {
       "epoch": 2.62,
-      "eval_accuracy": 0.6438356164383562,
-      "eval_loss": 1.6771588325500488,
-      "eval_runtime": 7.5025,
-      "eval_samples_per_second": 29.19,
-      "eval_steps_per_second": 3.732,
+      "eval_accuracy": 0.684931506849315,
+      "eval_loss": 0.8407436609268188,
+      "eval_runtime": 7.6113,
+      "eval_samples_per_second": 28.773,
+      "eval_steps_per_second": 3.679,
       "step": 325
     },
     {
       "epoch": 2.66,
-      "grad_norm": 4.947625160217285,
+      "grad_norm": 1.1653140783309937,
       "learning_rate": 6.733870967741935e-05,
-      "loss": 0.0964,
+      "loss": 0.3924,
       "step": 330
     },
     {
       "epoch": 2.74,
-      "grad_norm": 0.07066596299409866,
+      "grad_norm": 3.533656358718872,
       "learning_rate": 6.330645161290322e-05,
-      "loss": 0.0792,
+      "loss": 0.3786,
       "step": 340
     },
     {
       "epoch": 2.82,
-      "grad_norm": 6.108673572540283,
+      "grad_norm": 3.4866058826446533,
       "learning_rate": 5.9274193548387104e-05,
-      "loss": 0.0516,
+      "loss": 0.3433,
       "step": 350
     },
     {
       "epoch": 2.82,
-      "eval_accuracy": 0.680365296803653,
-      "eval_loss": 1.4380197525024414,
-      "eval_runtime": 8.9956,
-      "eval_samples_per_second": 24.345,
-      "eval_steps_per_second": 3.113,
+      "eval_accuracy": 0.6484018264840182,
+      "eval_loss": 0.969610333442688,
+      "eval_runtime": 8.4636,
+      "eval_samples_per_second": 25.875,
+      "eval_steps_per_second": 3.308,
       "step": 350
     },
     {
       "epoch": 2.9,
-      "grad_norm": 0.04943337291479111,
+      "grad_norm": 0.8572694659233093,
       "learning_rate": 5.5241935483870966e-05,
-      "loss": 0.024,
+      "loss": 0.3991,
       "step": 360
     },
     {
       "epoch": 2.98,
-      "grad_norm": 0.2530857026576996,
+      "grad_norm": 2.2343902587890625,
       "learning_rate": 5.120967741935484e-05,
-      "loss": 0.0136,
+      "loss": 0.2874,
       "step": 370
     },
     {
       "epoch": 3.02,
-      "eval_accuracy": 0.6484018264840182,
-      "eval_loss": 1.6561062335968018,
-      "eval_runtime": 7.9464,
-      "eval_samples_per_second": 27.56,
-      "eval_steps_per_second": 3.524,
+      "eval_accuracy": 0.680365296803653,
+      "eval_loss": 0.9290387034416199,
+      "eval_runtime": 7.9323,
+      "eval_samples_per_second": 27.608,
+      "eval_steps_per_second": 3.53,
       "step": 375
     },
     {
       "epoch": 3.06,
-      "grad_norm": 0.07669265568256378,
+      "grad_norm": 0.40827465057373047,
       "learning_rate": 4.7177419354838716e-05,
-      "loss": 0.0114,
+      "loss": 0.1757,
       "step": 380
     },
     {
       "epoch": 3.15,
-      "grad_norm": 0.04155677929520607,
+      "grad_norm": 0.28673678636550903,
       "learning_rate": 4.3145161290322584e-05,
-      "loss": 0.0109,
+      "loss": 0.122,
       "step": 390
     },
     {
       "epoch": 3.23,
-      "grad_norm": 0.8200215101242065,
+      "grad_norm": 1.3551135063171387,
       "learning_rate": 3.911290322580645e-05,
-      "loss": 0.0325,
+      "loss": 0.1682,
       "step": 400
     },
     {
       "epoch": 3.23,
-      "eval_accuracy": 0.6621004566210046,
-      "eval_loss": 1.6027804613113403,
-      "eval_runtime": 8.0216,
-      "eval_samples_per_second": 27.301,
-      "eval_steps_per_second": 3.491,
+      "eval_accuracy": 0.6575342465753424,
+      "eval_loss": 0.9713302254676819,
+      "eval_runtime": 7.4128,
+      "eval_samples_per_second": 29.543,
+      "eval_steps_per_second": 3.777,
       "step": 400
     },
     {
       "epoch": 3.31,
-      "grad_norm": 0.028287330642342567,
+      "grad_norm": 0.3541167676448822,
       "learning_rate": 3.508064516129033e-05,
-      "loss": 0.0085,
+      "loss": 0.1371,
       "step": 410
     },
     {
       "epoch": 3.39,
-      "grad_norm": 0.05001891031861305,
+      "grad_norm": 1.2407793998718262,
       "learning_rate": 3.1048387096774195e-05,
-      "loss": 0.0149,
+      "loss": 0.1575,
       "step": 420
     },
     {
       "epoch": 3.43,
-      "eval_accuracy": 0.6621004566210046,
-      "eval_loss": 1.6261166334152222,
-      "eval_runtime": 7.3858,
-      "eval_samples_per_second": 29.652,
-      "eval_steps_per_second": 3.791,
+      "eval_accuracy": 0.680365296803653,
+      "eval_loss": 0.9963440299034119,
+      "eval_runtime": 7.3009,
+      "eval_samples_per_second": 29.996,
+      "eval_steps_per_second": 3.835,
       "step": 425
     },
     {
       "epoch": 3.47,
-      "grad_norm": 0.07340113818645477,
+      "grad_norm": 2.0023341178894043,
       "learning_rate": 2.7016129032258064e-05,
-      "loss": 0.0104,
+      "loss": 0.1432,
       "step": 430
     },
     {
       "epoch": 3.55,
-      "grad_norm": 0.04165619611740112,
+      "grad_norm": 0.3904053568840027,
       "learning_rate": 2.2983870967741935e-05,
-      "loss": 0.0119,
+      "loss": 0.1827,
       "step": 440
     },
     {
       "epoch": 3.63,
-      "grad_norm": 0.032720282673835754,
+      "grad_norm": 0.42743560671806335,
       "learning_rate": 1.8951612903225807e-05,
-      "loss": 0.0082,
+      "loss": 0.0822,
       "step": 450
     },
     {
       "epoch": 3.63,
-      "eval_accuracy": 0.6621004566210046,
-      "eval_loss": 1.6614654064178467,
-      "eval_runtime": 7.1739,
-      "eval_samples_per_second": 30.527,
-      "eval_steps_per_second": 3.903,
+      "eval_accuracy": 0.7123287671232876,
+      "eval_loss": 0.9472664594650269,
+      "eval_runtime": 7.4585,
+      "eval_samples_per_second": 29.362,
+      "eval_steps_per_second": 3.754,
       "step": 450
     },
     {
       "epoch": 3.71,
-      "grad_norm": 0.07252111285924911,
+      "grad_norm": 0.7994292974472046,
       "learning_rate": 1.4919354838709679e-05,
-      "loss": 0.0079,
+      "loss": 0.133,
       "step": 460
     },
     {
       "epoch": 3.79,
-      "grad_norm": 0.029246076941490173,
+      "grad_norm": 0.4317198097705841,
       "learning_rate": 1.0887096774193549e-05,
-      "loss": 0.0093,
+      "loss": 0.1678,
       "step": 470
     },
     {
       "epoch": 3.83,
-      "eval_accuracy": 0.6529680365296804,
-      "eval_loss": 1.6878423690795898,
-      "eval_runtime": 7.9237,
-      "eval_samples_per_second": 27.639,
-      "eval_steps_per_second": 3.534,
+      "eval_accuracy": 0.7031963470319634,
+      "eval_loss": 0.9787965416908264,
+      "eval_runtime": 7.434,
+      "eval_samples_per_second": 29.459,
+      "eval_steps_per_second": 3.766,
       "step": 475
     },
     {
       "epoch": 3.87,
-      "grad_norm": 0.029909249395132065,
+      "grad_norm": 0.2247464805841446,
       "learning_rate": 6.854838709677419e-06,
-      "loss": 0.0087,
+      "loss": 0.148,
       "step": 480
     },
     {
       "epoch": 3.95,
-      "grad_norm": 0.042575616389513016,
+      "grad_norm": 0.28201037645339966,
       "learning_rate": 2.82258064516129e-06,
-      "loss": 0.0293,
+      "loss": 0.1406,
       "step": 490
     },
     {
       "epoch": 4.0,
       "step": 496,
       "total_flos": 6.10974224738132e+17,
-      "train_loss": 0.14545002643529686,
-      "train_runtime": 906.411,
-      "train_samples_per_second": 8.698,
-      "train_steps_per_second": 0.547
+      "train_loss": 0.601069396421794,
+      "train_runtime": 693.9158,
+      "train_samples_per_second": 11.362,
+      "train_steps_per_second": 0.715
     }
   ],
   "logging_steps": 10,