This project demonstrates how to perform real-time sentiment analysis using the distilbert-base-uncased-finetuned-sst-2-english transformer model from Hugging Face. It classifies input movie reviews as Positive or Negative, leveraging the power of pre-trained NLP models.
As an aspiring AI/ML software engineer, I wanted to explore how production-ready NLP models work under the hood. This project helped me understand how to:
- Load and fine-tune Hugging Face transformers
- Tokenize and prepare raw text for model inference
- Work with PyTorch tensors and GPU acceleration
- Deploy fast, lightweight NLP pipelines
I’m passionate about building impactful AI systems — whether it's sentiment analysis, medical imaging, or agentic AI. This is one of several projects where I'm diving deeper into language models and real-world inference pipelines.
- Model:
distilbert-base-uncased-finetuned-sst-2-english - Dataset: Stanford Sentiment Treebank (SST-2)
- Framework: PyTorch + Hugging Face Transformers
- Python
- PyTorch
- Hugging Face Transformers
- Google Colab (for development)
- GitHub (version control)
- Clone the repository:
git clone https://github.com/smartha2003/sentiment-analysis-bert.git
cd sentiment-analysis-bert