AI-Generated Poetry: Exploring Shakespearean Sonnets
Project Description
The AI-Generated Poetry project leverages deep learning techniques to generate poetry in the style of Shakespearean sonnets. Using a Long Short-Term Memory (LSTM) neural network trained on Shakespeare's sonnets, the model can predict and generate text based on a given seed text. The project involves preprocessing textual data, training the LSTM model with TensorFlow and Keras, and generating poetic verses using probabilistic sampling.
Role and Contributions
- Implemented the LSTM model architecture using TensorFlow and Keras.
- Preprocessed Shakespearean sonnets dataset for training.
- Developed the Python script for generating poetry based on user input.
- Configured and trained the LSTM model to optimize poetry generation.
Outcomes and Results
- Successfully generated coherent and stylistically accurate Shakespearean-like poetry.
- Explored different model configurations and hyperparameters to improve poetry quality.
- Shared the project on GitHub for transparency and further collaboration.
Technologies Used
- Python: Developed the LSTM model and script for text generation.
- TensorFlow / Keras: Used for implementing and training the LSTM neural network.
- NumPy: Handled numerical operations and data preprocessing.
- Git / GitHub: Version control and project management.
- JSON: Stored character mappings and configurations.
Challenges Faced and Solutions
- Challenge: Ensuring the LSTM model generates coherent and meaningful poetry.
Solution: Tuned model hyperparameters and utilized temperature sampling for varied text generation. - Challenge: Handling unseen characters and data preprocessing.
Solution: Implemented robust handling for character mappings and data vectorization.