Downloads · 30 days
9
27% of all-time downloads
Werea-TR-QA
GoktugD/Werea-TR-QA
Werea-TR-QA is a question answering model from GoktugD. Use it when the input is a question plus a passage. The card lists the license as mit.
All-time downloads
33
Public
Parameters
110M
440 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors440 MB · 100%
At a glance
- Task
- Question Answering
- License
- mit
- Model type
- bert
- Access
- Public
- Created
- Aug 21, 2026
- Updated
- Aug 21, 2026
- SHA
- 7b6bcd4a
Base models
Datasets
Model card
From the Hugging Face model README
ℹ️ Werea-co/Werea-TR-QA aynasıdır.
Werea-TR-QA — Türkçe Çıkarımsal Soru-Cevap
Verilen bağlamdan cevabı çıkaran (span) Türkçe QA modeli — RAG için ideal. Gerçek, insan-etiketli TQuAD2 ile eğitildi; iki tanınan gerçek benchmark'ta ölçüldü.
Benchmark (gerçek, insan-etiketli)
| Set | EM | F1 |
|---|---|---|
| TQuAD2 (validation) | 53.8% | 72.7% |
| XQuAD-tr | 46.1% | 64.5% |
BERTurk tabanlı, SQuAD-stili, 2 epoch.
Kullanım
from transformers import pipeline
qa=pipeline("question-answering",model="Werea-co/Werea-TR-QA")
print(qa(question="Werea nerede kuruldu?", context="Werea İstanbul'da kurulan bir yapay zeka şirketidir."))
💜 werea.co
- Task
- Question Answering
- Type
- bert
- License
- mit
- Languages
- tr
- Created
- Aug 21, 2026
- Updated
- Aug 21, 2026