update model card README.md

README.md

@@ -14,17 +14,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # polibert-malaysia-ver4
 
-This model is new version of YagiASAFAS/polibert-malaysia-ver3.
-What is new is that this model used a new dataset which not only used tnwei/ms-newspapers dataset but also almost 10k of instagram posts regarding several topics about Malaysia.
-By doing so, this model captures not only formal sentences such as News, but also captures informal sentences such as personal posts.
-As a tradeoff of ver3, the accuracy was quite lower compared to the previous one(ver2).
-This time we extracted data which have stronger characteristics regarding text features from the dataset used in ver3.
-By doing so, now this model captures not only formal sentences such as News, but also captures informal sentences such as personal posts.
-And also the accuracy is higher than the previous one(ver).
-However, several data were deleted from the dataset. Thus the devirsity has decreased. This is the tradeoff for this time.
+This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3164
-- Accuracy: 0.9536
+- Loss: 0.2641
+- Accuracy: 0.9371
 
 ## Model description
 
@@ -45,38 +38,25 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 3e-05
 - train_batch_size: 8
-- eval_batch_size: 64
+- eval_batch_size: 8
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 8
 - mixed_precision_training: Native AMP
 
-### Label Mappings
-- 0: Economic Concerns
-- 1: Racial discrimination or polarization
-- 2: Leadership weaknesses
-- 3: Development and infrastructure gaps
-- 4: Corruption
-- 5: Political instablility
-- 6: Socials and Public safety
-- 7: Administration
-- 8: Education
-- 9: Religion issues
-- 10: Environmental
-
 ### Training results
 
-| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
-|:-------------:|:-----:|:-----:|:---------------:|:--------:|
-| 0.4227        | 1.0   | 1250  | 0.4466          | 0.914    |
-| 0.2549        | 2.0   | 2500  | 0.3277          | 0.9396   |
-| 0.1714        | 3.0   | 3750  | 0.3590          | 0.9424   |
-| 0.1298        | 4.0   | 5000  | 0.3354          | 0.946    |
-| 0.1002        | 5.0   | 6250  | 0.3634          | 0.9428   |
-| 0.0691        | 6.0   | 7500  | 0.3164          | 0.9536   |
-| 0.0845        | 7.0   | 8750  | 0.3469          | 0.9488   |
-| 0.0527        | 8.0   | 10000 | 0.3535          | 0.9488   |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|
+| 1.032         | 1.0   | 938  | 0.3282          | 0.9227   |
+| 0.2967        | 2.0   | 1876 | 0.2641          | 0.9371   |
+| 0.1987        | 3.0   | 2814 | 0.2902          | 0.9403   |
+| 0.163         | 4.0   | 3752 | 0.2995          | 0.9451   |
+| 0.1315        | 5.0   | 4690 | 0.2922          | 0.9445   |
+| 0.0864        | 6.0   | 5628 | 0.2760          | 0.9504   |
+| 0.0861        | 7.0   | 6566 | 0.2836          | 0.9493   |
+| 0.0686        | 8.0   | 7504 | 0.2933          | 0.9509   |
 
 
 ### Framework versions