Views
No views yet
distilbert-base-multilingual-cased1: End of Utterance (EOU)0: Not End of Utterance (NOT_EOU)hindi-conversational-eou dataset — a balanced collection of 1000 Hindi conversational phrases labeled for end-of-turn detection."text": The utterance string"label": 0 or 1 (as defined above)(Note: These are example metrics — replace with your actual numbers if available)
1from transformers import pipeline
2
3classifier = pipeline("text-classification", model="yashsoni78/distilbert-hindi-eou-detector")
4
5# Example phrases
6examples = [
7 "क्या तुम मेरे साथ चलोगे?",
8 "अगर हम वहाँ जाते तो",
9]
10
11for text in examples:
12 result = classifier(text)
13 print(f"{text} => {result}")
14
15🔍 Limitations
16 - Trained on a small dataset (1000 examples); may not generalize to complex or domain-specific Hindi.
17 - Only binary EOU detection, no deeper semantic understanding.
18 - Assumes input is in colloquial conversational Hindi.
19
20🧾 Citation
21If you use this model in your research or application, please cite:
22
23@misc{distilbert_hindi_eou_2025,
24 title = {distilbert-hindi-eou-detector},
25 author = {Yash Soni},
26 year = {2025},
27 howpublished = {\url{https://huggingface.co/yashsoni78/distilbert-hindi-eou-detector}},
28 note = {Fine-tuned model for Hindi end-of-utterance detection}
29}
30
31📄 License
32This model is released under the MIT License. You are free to use, modify, and distribute with attribution.
33
34🙏 Acknowledgements
35 - Base model: distilbert-base-multilingual-cased
36 - Dataset: hindi-end-of-utterance-detection
37 - Created with the help of 🤗 Transformers