Sentimaster is a web-based sentiment analysis tool developed to analyze restaurant reviews using AI-powered sentiment classification. Built using a fine-tuned GPT-2 model, Sentimaster empowers businesses with real-time actionable insights from customer feedback to improve customer service and decision-making.
- GPT-2 (124M parameters) for sentiment analysis model
- Flask for backend web framework
- AWS EC2 for deployment
- HTML, CSS, JavaScript for frontend development
- PyTorch for model training and inference
- Tiktoken for tokenization
- Sentiment Analysis: Classifies restaurant reviews as either "Positive" or "Negative".
- Real-Time Feedback: Users can submit reviews through the web interface and receive real-time sentiment analysis.
- Model: Fine-tuned GPT-2 (124M parameters) on restaurant-specific data for improved accuracy.
- Deployment: Deployed using Flask, accessible via a web browser.
-
Clone this repository to your local machine:
git clone https://github.com/GoldSharon/Sentimaster.git
-
Navigate to the project directory:
cd Sentimaster -
Install the required dependencies:
pip install -r requirements.txt
-
Ensure that you have the pretrained model weights (
model_and_optimizer.pth) placed in the project directory.- If the model weights are not available, follow the training guidelines provided in the documentation or scripts to train the model.
-
Run the Flask app:
python app.py
-
The application will be available at:
http://127.0.0.1:5000/ -
Open your browser and go to the URL. Enter a restaurant review in the text field, and the model will classify the sentiment as "Positive" or "Negative".
- GPT-2 (124M parameters) is used for sentiment analysis, fine-tuned on restaurant-related reviews to improve domain-specific accuracy.
- Architecture Details:
- 12 layers, 768 embedding dimensions, and 12 attention heads.
- Trained and optimized using the Adam optimizer with a learning rate of 0.0004.
/: Home route, renders the input form for the review./submit: POST method, accepts the review and returns sentiment classification./result: Displays the result of the sentiment analysis.
- The user submits a restaurant review via the form.
- The model classifies the sentiment of the review (Positive/Negative).
- The result is displayed on a new page, showing whether the review is positive or negative.
By using Sentimaster, businesses can:
- Gain insights from customer feedback.
- Improve customer service and satisfaction.
- Make data-driven decisions for business growth and improvement.
Feel free to fork the repository, create issues, and submit pull requests. Contributions are always welcome!
This project is licensed under the MIT License.