model details
Browse files
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
This model is part of the Research topic "Bias and Fairness in AI" conducted by Shaina Raza, Deepak John Reji, Chen Ding
|
| 2 |
+
|
| 3 |
+
- language: en
|
| 4 |
+
- tags: Token Classification
|
| 5 |
+
- Dataset : MBAD Data
|
| 6 |
+
- datasize : 17775 entries
|
| 7 |
+
- Label distribution : Biased - 10651, Non-Biased - 7124
|
| 8 |
+
- Train-Test split : 90 : 10
|
| 9 |
+
- tokenizer : distilbert-base-uncased
|
| 10 |
+
- model : distilbert-base-uncased
|
| 11 |
+
- optimizer : adam
|
| 12 |
+
- Learning rate : 5e-5
|
| 13 |
+
- Model parameters : 66,955,010
|
| 14 |
+
- epochs : 30
|
| 15 |
+
- Train accuracy : 0.7697
|
| 16 |
+
- Test accuracy : 0.62
|
| 17 |
+
- Train loss : 0.4506
|
| 18 |
+
- Test loss : 0.9644
|
| 19 |
+
- Carbon emission 0.319355 Kg
|
| 20 |
+
|