|
Download README.md from Werea-co/Werea-TR-Topic: direct link, hf CLI and curl.
- Browser
- Download file 830 Bytes
-
https://huggingface.co/Werea-co/Werea-TR-Topic/resolve/main/README.md
- Command line
-
hf download hf://Werea-co/Werea-TR-Topic/README.md
-
curl -L -o README.md https://huggingface.co/Werea-co/Werea-TR-Topic/resolve/main/README.md
830 Bytes
metadata
language:
- tr
license: mit
pipeline_tag: text-classification
base_model: dbmdz/bert-base-turkish-cased
tags:
- turkish
- text-classification
- werea
datasets:
- savasy/ttc4900
metrics:
- accuracy
- f1
Werea-TR-Topic — Türkçe haber konu sınıflandırma
Gerçek, insan-etiketli savasy/ttc4900 verisiyle eğitildi; aynı setin ayrılmış test bölümünde dürüstçe ölçüldü. 7 sınıf.
Benchmark (gerçek test)
| Metrik | Skor |
|---|---|
| Accuracy | 92.8% |
| Macro-F1 | 92.7% |
BERTurk tabanlı, 6 epoch.
Kullanım
from transformers import pipeline
clf=pipeline("text-classification",model="Werea-co/Werea-TR-Topic")
print(clf("text metni"))
💜 werea.co