TinySentimentClassifier

Overview

TinySentimentClassifier is a compact BERT-based model fine-tuned for sentiment analysis on English text. It classifies input text into three categories: positive, neutral, or negative. Designed for efficiency, it is suitable for deployment on resource-constrained environments while maintaining strong performance on standard sentiment datasets.

Model Architecture

• Base model: DistilBERT (distilled version of BERT-base-uncased)
• Task head: Sequence classification head with 3 output labels
• Hidden size: 768
• Number of layers: 6
• Parameters: ~66M

The model follows the standard BertForSequenceClassification architecture from the Transformers library.

Usage

from transformers import pipeline

classifier = pipeline(
    "sentiment-analysis",
    model="your-username/TinySentimentClassifier",
    return_all_scores=False
)

result = classifier("I love this product!")
print(result)
# [{'label': 'positive', 'score': 0.99}]