Arabic EOU Detection Model
Fine-tuned aubmindlab/bert-base-arabertv2 for binary end-of-utterance detection
in Saudi Arabic conversational speech.
Labels
- 0: Not end of utterance
- 1: End of utterance
Intended Use
Real-time conversational agents and turn-taking systems.
Example
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
tokenizer = AutoTokenizer.from_pretrained("YOUR_USERNAME/arabic-eou-bertv2")
model = AutoModelForSequenceClassification.from_pretrained("YOUR_USERNAME/arabic-eou-bertv2")
inputs = tokenizer("تمام خلاص", return_tensors="pt")
prob = torch.softmax(model(**inputs).logits, dim=-1)[0,1].item()