WideMan commited on
Commit
c28284c
1 Parent(s): 2764067

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -15
README.md CHANGED
@@ -15,25 +15,15 @@ bass-model: GroNLP/bert-base-dutch-cased
15
  This model is a fine-tuned version of [GroNLP/bert-base-dutch-cased](https://huggingface.co/GroNLP/bert-base-dutch-cased) on [GroNLP/dutch-cola](https://huggingface.co/datasets/GroNLP/dutch-cola).
16
  It achieves the following results on the evaluation set:
17
 
18
- - Loss: 0.5693
19
- - Accuracy: 0.7838
20
 
21
 
22
  #### Training Hyperparameters
23
 
24
  The following hyperparameters were used during training:
25
 
26
- - learning rate: 2e-05
27
- - train_batch_size: 16
28
- - eval_batch_size: 16
29
  - seed: 42
30
- - num_epochs=3
31
-
32
-
33
- ### Training results
34
-
35
- | Training Loss | Epoch | Step | Validation Loss | Accuracy |
36
- |:-------------:|:-----:|:----:|:---------------:|:--------:|
37
- | 0.5233 | 1.0 | 1244 | 0.5762 | 0.7396 |
38
- | 0.2903 | 2.0 | 2488 | 0.5693 | 0.7838 |
39
- | 0.1699 | 3.0 | 3732 | 0.8232 | 0.7846 |
 
15
  This model is a fine-tuned version of [GroNLP/bert-base-dutch-cased](https://huggingface.co/GroNLP/bert-base-dutch-cased) on [GroNLP/dutch-cola](https://huggingface.co/datasets/GroNLP/dutch-cola).
16
  It achieves the following results on the evaluation set:
17
 
18
+ - Accuracy: 0.7938
 
19
 
20
 
21
  #### Training Hyperparameters
22
 
23
  The following hyperparameters were used during training:
24
 
25
+ - learning rate: 5e-05
26
+ - train_batch_size: 32
27
+ - eval_batch_size: 32
28
  - seed: 42
29
+ - num_epochs=2