Recurrent Neural Network (RNN) for Text Generation

Overview

This project demonstrates how to build and train a Recurrent Neural Network (RNN) using TensorFlow and Keras for generating text. The model is designed to predict the next word in a sequence based on a given input text.

Components

1. Data Loading and Preprocessing

   • Data is loaded from a CSV file into a Pandas DataFrame.
   • The DataFrame is converted into a list of lists and flattened into a single list of sentences.
   • Text data is tokenized using TensorFlow's Tokenizer to convert words into integer sequences.
   • Sequences are padded to ensure uniform length.

2. Model Architecture

   • Embedding Layer: Converts integer sequences into dense vectors of a fixed size.
   • Simple RNN Layer: Processes the sequence data and retains temporal information.
   • Dense Layer: Outputs a probability distribution over the vocabulary using a softmax activation function.

3. Model Training

   • The data is split into features (input sequences) and targets (next words).
   • The model is compiled with the Adam optimizer and sparse categorical crossentropy loss function.
   • Training is conducted over 50 epochs with a validation split to monitor performance.

4. Prediction

   • A function is provided to generate predictions of the next word(s) given a seed text.
   • The function tokenizes and pads the seed text, uses the model to predict the next word, and updates the seed text with the predicted word.

Results

• The model's performance is evaluated using accuracy and loss metrics. The trained model can generate text based on the provided seed text.

Usage

• Ensure you have the required libraries installed (TensorFlow, NumPy, pandas).
• Prepare your text data in a CSV file named data.csv.
• Run the code to train the model and generate text based on a seed input.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For any questions or feedback, please reach out to SURESH BEEKHANI