File size: 1,188 Bytes
193f60b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 | ---
language: ar
license: apache-2.0
tags:
- text-classification
- arabic
- arabert
pipeline_tag: text-classification
---
# 7oudaModel - AraBERT Call Classifier
موديل مبني على [aubmindlab/bert-base-arabertv02](https://huggingface.co/aubmindlab/bert-base-arabertv02)
اتعمله Fine-tuning لتصنيف إفادات المكالمات إلى **ناجحة** أو **غير ناجحة**.
## الاستخدام
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
tokenizer = AutoTokenizer.from_pretrained("Mahmoud252002/7oudaModel")
model = AutoModelForSequenceClassification.from_pretrained("Mahmoud252002/7oudaModel")
text = "ضع نص الإفادة هنا"
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=256)
with torch.no_grad():
logits = model(**inputs).logits
pred = torch.argmax(logits, dim=1).item()
print("ناجحة" if pred == 1 else "غير ناجحة")
```
## تفاصيل التدريب
- الموديل الأساسي: `aubmindlab/bert-base-arabertv02`
- المهمة: تصنيف ثنائي (Binary Classification)
- عدد الفئات: 2 (ناجحة / غير ناجحة) |