Keras Sentiment Analysis Model -- 2
Budget: $8 – $15 USD
I need a Keras-based sentiment analysis model built from my existing labeled dataset. You will:
• Inspect and preprocess the text (tokenization, padding, handling out-of-vocabulary words).
• Design an appropriate neural network in Keras—an LSTM, GRU, CNN, or a hybrid architecture that you feel best fits the data.
• Train, validate, and fine-tune the model, tracking accuracy, precision, recall, and F1.
• Provide clean, well-commented Python code plus a brief README that explains setup, training, and how to make predictions on new text.
• Hand over the trained model weights and any scripts/notebooks used.
I’m looking for clear, reproducible work that I can continue to build on, delivered through GitHub or a shared drive.
• Inspect and preprocess the text (tokenization, padding, handling out-of-vocabulary words).
• Design an appropriate neural network in Keras—an LSTM, GRU, CNN, or a hybrid architecture that you feel best fits the data.
• Train, validate, and fine-tune the model, tracking accuracy, precision, recall, and F1.
• Provide clean, well-commented Python code plus a brief README that explains setup, training, and how to make predictions on new text.
• Hand over the trained model weights and any scripts/notebooks used.
I’m looking for clear, reproducible work that I can continue to build on, delivered through GitHub or a shared drive.