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 [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2821
-- Accuracy: 0.9034
+- Loss: 0.2854
+- Accuracy: 0.9055
 
 ## Model description
 
@@ -38,30 +38,40 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-05
-- train_batch_size: 16
-- eval_batch_size: 16
+- train_batch_size: 32
+- eval_batch_size: 32
 - seed: 42
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 64
+- total_train_batch_size: 128
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 10
+- num_epochs: 20
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 0.5334        | 1.0   | 750  | 0.5595          | 0.8302   |
-| 0.3389        | 2.0   | 1500 | 0.4090          | 0.8602   |
-| 0.358         | 3.0   | 2250 | 0.3631          | 0.8717   |
-| 0.3672        | 4.0   | 3000 | 0.3368          | 0.8815   |
-| 0.3458        | 5.0   | 3750 | 0.3231          | 0.8842   |
-| 0.2721        | 6.0   | 4500 | 0.3075          | 0.8885   |
-| 0.2397        | 7.0   | 5250 | 0.3035          | 0.8899   |
-| 0.2779        | 8.0   | 6000 | 0.2893          | 0.8963   |
-| 0.2046        | 9.0   | 6750 | 0.2868          | 0.8991   |
-| 0.2599        | 10.0  | 7500 | 0.2821          | 0.9034   |
+| 0.7694        | 1.0   | 375  | 0.7344          | 0.8202   |
+| 0.4737        | 2.0   | 750  | 0.5042          | 0.8457   |
+| 0.4318        | 3.0   | 1125 | 0.3876          | 0.8702   |
+| 0.3395        | 4.0   | 1500 | 0.3688          | 0.8769   |
+| 0.3105        | 5.0   | 1875 | 0.3357          | 0.8845   |
+| 0.2742        | 6.0   | 2250 | 0.3272          | 0.883    |
+| 0.2898        | 7.0   | 2625 | 0.3156          | 0.8903   |
+| 0.2774        | 8.0   | 3000 | 0.3004          | 0.8937   |
+| 0.2833        | 9.0   | 3375 | 0.2976          | 0.8933   |
+| 0.2398        | 10.0  | 3750 | 0.2954          | 0.8954   |
+| 0.2143        | 11.0  | 4125 | 0.2724          | 0.9055   |
+| 0.1808        | 12.0  | 4500 | 0.2843          | 0.8985   |
+| 0.2298        | 13.0  | 4875 | 0.2918          | 0.8968   |
+| 0.218         | 14.0  | 5250 | 0.2742          | 0.9036   |
+| 0.1885        | 15.0  | 5625 | 0.2932          | 0.8976   |
+| 0.1927        | 16.0  | 6000 | 0.2875          | 0.904    |
+| 0.1546        | 17.0  | 6375 | 0.2832          | 0.9066   |
+| 0.186         | 18.0  | 6750 | 0.2796          | 0.9054   |
+| 0.1515        | 19.0  | 7125 | 0.2850          | 0.9018   |
+| 0.1766        | 20.0  | 7500 | 0.2854          | 0.9055   |
 
 
 ### Framework versions