Baps24/odia-sentiment-muril-v2

Odia Sentiment Classifier v2

Fine-tuned google/muril-base-cased for Odia language sentiment analysis.

Built with ❤️ for the 45 million Odia speakers of Odisha 🇮🇳

Dedicated to Lord Jagannath and the people of Odisha 🙏

Improvements over v1

• Larger dataset: 66 samples (auto-labeled from OdiaGenAI corpus)
• Better class balance
• Higher accuracy

Usage

from transformers import pipeline
pipe = pipeline("text-classification", model="Baps24/odia-sentiment-muril-v2")
pipe("ଏହି ଉତ୍ପାଦ ଭଲ")  # Positive
pipe("ଏହି ଜିନିଷ ଖରାପ")  # Negative

Labels

• Positive (ସକାରାତ୍ମକ)
• Negative (ନକାରାତ୍ମକ)
• Neutral (ନିରପେକ୍ଷ)

Training

• Base: google/muril-base-cased
• Dataset: Baps24/odia-sentiment-dataset
• Accuracy: 0.300
• F1 Score: 0.138

Community

This model is part of the Odias in ML initiative. Visit: https://odisha-ml.github.io