factual-consistency-classification-ja

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [line-corporation/line-distilbert-base-japanese](https://huggingface.co/line-corporation/line-distilbert-base-japanese) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8124
-- Accuracy: 0.6777
+- Loss: 0.9166
+- Accuracy: 0.4707
 
 ## Model description
 
@@ -44,17 +44,22 @@ The following hyperparameters were used during training:
 - distributed_type: tpu
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 5
+- num_epochs: 10
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| No log        | 1.0   | 306  | 0.8936          | 0.5762   |
-| 0.9294        | 2.0   | 612  | 0.8541          | 0.6367   |
-| 0.9294        | 3.0   | 918  | 0.8295          | 0.6621   |
-| 0.8671        | 4.0   | 1224 | 0.8167          | 0.6777   |
-| 0.8414        | 5.0   | 1530 | 0.8124          | 0.6777   |
+| No log        | 1.0   | 306  | 1.0543          | 0.2891   |
+| 1.0487        | 2.0   | 612  | 1.0134          | 0.3301   |
+| 1.0487        | 3.0   | 918  | 0.9846          | 0.3457   |
+| 0.9961        | 4.0   | 1224 | 0.9620          | 0.3965   |
+| 0.963         | 5.0   | 1530 | 0.9462          | 0.4258   |
+| 0.963         | 6.0   | 1836 | 0.9326          | 0.4688   |
+| 0.9401        | 7.0   | 2142 | 0.9254          | 0.4707   |
+| 0.9401        | 8.0   | 2448 | 0.9208          | 0.4688   |
+| 0.9295        | 9.0   | 2754 | 0.9176          | 0.4707   |
+| 0.9232        | 10.0  | 3060 | 0.9166          | 0.4707   |
 
 
 ### Framework versions