update model card README.md

README.md

@@ -17,11 +17,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [aubmindlab/araelectra-base-discriminator](https://huggingface.co/aubmindlab/araelectra-base-discriminator) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1577
-- Macro F1: 0.9379
-- Accuracy: 0.9398
-- Precision: 0.9372
-- Recall: 0.9386
+- Loss: 0.2410
+- Macro F1: 0.9174
+- Accuracy: 0.9197
+- Precision: 0.9159
+- Recall: 0.9191
 
 ## Model description
 
@@ -40,23 +40,24 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-05
+- learning_rate: 3e-05
 - train_batch_size: 16
-- eval_batch_size: 32
+- eval_batch_size: 16
 - seed: 123
 - gradient_accumulation_steps: 2
 - total_train_batch_size: 32
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 3
+- num_epochs: 4
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Macro F1 | Accuracy | Precision | Recall |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:---------:|:------:|
-| 0.3524        | 1.0   | 798  | 0.1748          | 0.9312   | 0.9335   | 0.9324    | 0.9300 |
-| 0.2299        | 2.0   | 1596 | 0.1678          | 0.9316   | 0.9340   | 0.9330    | 0.9303 |
-| 0.172         | 3.0   | 2394 | 0.1577          | 0.9379   | 0.9398   | 0.9372    | 0.9386 |
+| 0.1488        | 1.0   | 798  | 0.2533          | 0.9109   | 0.9125   | 0.9078    | 0.9179 |
+| 0.1264        | 2.0   | 1596 | 0.2410          | 0.9174   | 0.9197   | 0.9159    | 0.9191 |
+| 0.145         | 3.0   | 2394 | 0.2779          | 0.9083   | 0.9117   | 0.9107    | 0.9064 |
+| 0.0934        | 4.0   | 3192 | 0.2858          | 0.9171   | 0.9199   | 0.9179    | 0.9163 |
 
 
 ### Framework versions