update model card README.md

README.md

@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 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 imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1475
-- Accuracy: 0.9792
+- Loss: 0.0581
+- Accuracy: 0.9904
 
 ## Model description
 
@@ -44,8 +44,8 @@ The following hyperparameters were used during training:
 - seed: 42
 - distributed_type: IPU
 - gradient_accumulation_steps: 32
-- total_train_batch_size: 128
-- total_eval_batch_size: 16
+- total_train_batch_size: 32
+- total_eval_batch_size: 4
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_ratio: 0.1
@@ -56,9 +56,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 0.3961        | 1.0   | 189  | 0.2964          | 0.9609   |
-| 0.3819        | 2.0   | 378  | 0.1704          | 0.9773   |
-| 0.152         | 3.0   | 567  | 0.1475          | 0.9792   |
+| 0.0804        | 1.0   | 759  | 0.1383          | 0.9741   |
+| 0.0385        | 2.0   | 1518 | 0.0756          | 0.9859   |
+| 0.1211        | 3.0   | 2277 | 0.0581          | 0.9904   |
 
 
 ### Framework versions