🌟 Onubhuti: Bangla Sentiment Analysis Model

Onubhuti is a sentiment analysis model specifically trained to classify Bangla text into Positive, Neutral, or Negative sentiments. Fine-tuned from FB XLM-Roberta, it leverages state-of-the-art NLP techniques for robust performance. This model can be applied to various natural language processing (NLP) tasks such as customer reviews, social media analysis, and more.

📖 Model Details

📋 Model Description

Onubhuti is a transformer-based model fine-tuned on Bangla text for sentiment classification tasks. It has been trained to classify Bangla sentences into three sentiment categories:

Positive: Sentiments that express approval, happiness, or general positivity.
Neutral: Sentiments that are neutral or have no clear positive or negative sentiment.
Negative: Sentiments that express disapproval, dissatisfaction, or general negativity.

The model uses FB XLM-Roberta, a transformer model pre-trained on multilingual data and further fine-tuned specifically for Bangla. This enables the model to understand nuances in sentiment for Bangla text with high accuracy.

Developed by: Nazmus Sakib Apurba
Funded by: Not Funded
Shared by: Tamim
Model type: Transformer-based classification model
Language(s): Bangla (বাংলা)
License: Apache 2.0 (or any relevant license)
Finetuned from model: FB XLM-Roberta

🔗 Model Sources

Repository: Your Hugging Face model URL
Paper: [Optional, link to relevant paper if available]
Demo: [Optional, link to any demo or related page]

🚀 Uses

Onubhuti is designed for sentiment analysis tasks and can be used in various contexts:

Social Media Monitoring: Analyzing tweets, Facebook posts, or other social media content for sentiment.
Customer Feedback Analysis: Classifying customer reviews, survey responses, or feedback.
Content Moderation: Identifying positive, negative, or neutral comments in a user-generated content environment.

🛠️ Installation and Usage

Installation:
To use this model, you need to install the necessary libraries:

pip install transformers torch

# Load model directly
from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("Apurba3036/onubhuti-bangla-sentiment")
model = AutoModelForSequenceClassification.from_pretrained("Apurba3036/onubhuti-bangla-sentiment")

💡 Example Output:

Bangla text: এই জায়গাটা এত সুন্দর যে ভাষায় বর্ণনা করা কঠিন। পাহাড় আর সবুজ গাছপালার মিশেলে প্রকৃতি যেন তার সেরা রূপ এখানে মেলে ধরেছে। আমি এই জায়গায় বারবার আসতে চাই।
Predicted sentiment: Positive