End of training

README.md

@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [kykim/bert-kor-base](https://huggingface.co/kykim/bert-kor-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2414
-- Accuracy: 0.9149
+- Loss: 0.5054
+- Accuracy: 0.8618
 
 ## Model description
 
@@ -37,25 +37,26 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 2e-05
-- train_batch_size: 128
-- eval_batch_size: 128
+- train_batch_size: 32
+- eval_batch_size: 32
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 2
+- num_epochs: 5
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| No log        | 0.23  | 100  | 0.3599          | 0.8855   |
-| No log        | 0.46  | 200  | 0.3003          | 0.8995   |
-| No log        | 0.7   | 300  | 0.2731          | 0.9092   |
-| No log        | 0.93  | 400  | 0.2682          | 0.9103   |
-| 0.3663        | 1.16  | 500  | 0.2616          | 0.9105   |
-| 0.3663        | 1.39  | 600  | 0.2522          | 0.9144   |
-| 0.3663        | 1.62  | 700  | 0.2462          | 0.9149   |
-| 0.3663        | 1.86  | 800  | 0.2414          | 0.9149   |
+| No log        | 0.55  | 100  | 0.4919          | 0.8230   |
+| No log        | 1.1   | 200  | 0.4253          | 0.8509   |
+| No log        | 1.65  | 300  | 0.4203          | 0.8556   |
+| No log        | 2.2   | 400  | 0.4475          | 0.8711   |
+| 0.411         | 2.75  | 500  | 0.4450          | 0.8649   |
+| 0.411         | 3.3   | 600  | 0.4582          | 0.8665   |
+| 0.411         | 3.85  | 700  | 0.4824          | 0.8649   |
+| 0.411         | 4.4   | 800  | 0.4810          | 0.8696   |
+| 0.411         | 4.95  | 900  | 0.5054          | 0.8618   |
 
 
 ### Framework versions