Fine-tuned microsoft/Multilingual-MiniLM-L12-H384 for three-class sentiment classification of Pakistani news coverage of the Aurat March feminist movement.

Labels: Oppositional · Supportive · Neutral

Developed as part of the Master's Thesis "Compact Transformer-Based Sentiment Analysis of Aurat March Media Coverage in Urdu and English" at HSE University, Nizhny Novgorod.

Usage

python

from transformers import pipeline

classifier = pipeline(
    "sentiment-analysis",
    model="aakashMeghwar01/aurat-march-sentiment-minilm"
)

# English example
classifier("The Aurat March is an important movement for women's rights in Pakistan")
# → [{'label': 'Supportive', 'score': 0.87}]

# Urdu example
classifier("عورت مارچ کے نعرے اسلامی اقدار کے خلاف ہیں")
# → [{'label': 'Oppositional', 'score': 0.79}]

Evaluation

Class	Precision	Recall	F1	Support
Oppositional	0.89	0.81	0.85	21
Supportive	0.83	0.88	0.86	17
Neutral	0.67	0.75	0.71	8
Weighted avg	0.83	0.83	0.83	46

Training

Base model: microsoft/Multilingual-MiniLM-L12-H384
Training corpus: 470 annotated Urdu and English news articles (Aurat March coverage 2018–2023)
Class weighting: WeightedTrainer with inverse-frequency weights to handle class imbalance
Epochs: 10 · Batch size: 16 · Learning rate: 2e-5
Hardware: Google Colab T4 GPU

Labels

ID	Label	Description
0	Oppositional	Article frames the Aurat March negatively
1	Supportive	Article frames the Aurat March positively
2	Neutral	Procedural/balanced reporting

Citation

Meghwar, A. (2025). Compact Transformer-Based Sentiment Analysis of Aurat March 
Media Coverage in Urdu and English. MA Thesis, HSE University Nizhny Novgorod.
Model: https://huggingface.co/aakashMeghwar01/aurat-march-sentiment-minilm

Contact

Aakash Meghwar · aakashmeghwar01@gmail.com · LinkedIn It utilizes a Compact Transformer (MiniLM-L12-v2) architecture, fine-tuned specifically to navigate the nuanced socio-political discourse of the Aurat March in Pakistan. The model is optimized for high performance on consumer-grade hardware (T4 GPUs), adhering to the "Green AI" philosophy of accessible NLP for low-resource languages.

🚀 About the Author

I am Aakash Meghwar, a Computational Linguist specializing in South Asian languages.

🎓 M.S. in Applied Linguistics & Text Analytics (HSE, Russia - Graduating June 2026)
🎓 B.S. in English Language & Literature (NUML, Islamabad)
💡 Founder: Text Tech Solutions
✍️ Published Researcher: Author of "Compact Transformer Models for Classical Urdu Poetry" (Corporum Journal).

🤝 Open for Collaboration & PhD Opportunities

I am actively seeking PhD opportunities and Research Collaborations in the following areas:

Low-Resource NLP: Developing efficient models for Sindhi, Urdu, and Siraiki.
Literary Informatics: Computational stylistics and affective registers in classical poetry.
SindhiLM Evolution: I am currently developing SindhiLM-v2 (featuring an improved Sindhi-BPE tokenizer and knowledge distillation).

Are you looking for a researcher to join your NLP lab or organization? I am open to discussing projects involving Model Compression, Cross-Lingual Transfer, and South Asian Language Tech.

📩 Contact & Links

Email: aakashmeghwar01@gmail.com
LinkedIn: Aakash Meghwar
Projects: SindhiLM | Urdu Poetry Research

How to use

python

from transformers import pipeline
classifier = pipeline("sentiment-analysis", model="aakashMeghwar01/aurat-march-sentiment-minilm")
# Example Urdu input
classifier("عورت مارچ خواتین کے حقوق کی ایک اہم تحریک ہے")