Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,39 @@
|
|
| 1 |
---
|
| 2 |
base_model: dbmdz/bert-base-turkish-uncased
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
base_model: dbmdz/bert-base-turkish-uncased
|
| 3 |
+
language:
|
| 4 |
+
- tr
|
| 5 |
+
library_name: transformers
|
| 6 |
---
|
| 7 |
+
* We use the trt turkish news data which inside news context and categories belongs to one of the news.
|
| 8 |
+
|
| 9 |
+
* Using bert base turkish uncased model, aimed to label the categories to the news.
|
| 10 |
+
|
| 11 |
+
* We have 11 separate categories as below;
|
| 12 |
+
|
| 13 |
+
('bilim_teknoloji',
|
| 14 |
+
|
| 15 |
+
'dunya', 'egitim',
|
| 16 |
+
|
| 17 |
+
'ekonomi',
|
| 18 |
+
|
| 19 |
+
'guncel',
|
| 20 |
+
|
| 21 |
+
'gundem',
|
| 22 |
+
|
| 23 |
+
'kultur_sanat',
|
| 24 |
+
|
| 25 |
+
'saglik',
|
| 26 |
+
|
| 27 |
+
'spor',
|
| 28 |
+
|
| 29 |
+
'turkiye',
|
| 30 |
+
|
| 31 |
+
'yasam')
|
| 32 |
+
|
| 33 |
+
* We got the validation skor and follow the metric accuracy. The model gave us successfully result.
|
| 34 |
+
|
| 35 |
+
* loss accuracy val_loss val_accuracy
|
| 36 |
+
0 0.739859 0.766797 0.507217 0.828693
|
| 37 |
+
1 0.413323 0.865625 0.474160 0.843466
|
| 38 |
+
|
| 39 |
+
|