Next Word Prediction with LSTM

This project demonstrates how to create a next word prediction model using an LSTM (Long Short-Term Memory) neural network. The model is trained on a given corpus of text data and predicts the next word in a sequence. The project uses TensorFlow and Keras for building the model, and Streamlit for creating a simple web interface to interact with the model.

Table of Contents

• Installation
• Usage
• Model Training
• Running the App
• Contributing
• License

Installation

1. Clone the repository:

   git clone https://github.com/Sahilkumar19/NEXT-WORD-PREDICTOR.git
   cd NEXT-WORD-PREDICTOR

2. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate   # On Windows use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

Usage

Model Training

1. Use hamlet.txt to train the model I have downloaded this from NLTK itself.

2. Run the training script:

   python experiments.py

   This script will preprocess the data, train the LSTM model, and save the trained model (next_word_lstm.h5) and tokenizer (tokenizer.pickle) to disk.

Running the App

1. Ensure the model and tokenizer files (next_word_lstm.h5 and tokenizer.pickle) are in the root directory.

2. Run the Streamlit app:

   streamlit run app.py

   This will start a web server and open a web interface where you can input a sequence of words and get the predicted next word.

Model Training

The model training process involves the following steps:

1. Data Preprocessing: The text data is tokenized and converted into sequences of word indices. These sequences are then padded to ensure uniform length.

2. Model Definition: An LSTM model is defined using Keras, consisting of embedding, LSTM, dropout, and dense layers.

3. Model Training: The model is trained on the preprocessed data using categorical cross-entropy loss and the Adam optimizer.

4. Model Saving: The trained model and tokenizer are saved to disk for later use in prediction.

Running the App

The Streamlit app provides a simple web interface for interacting with the trained model. Users can input a sequence of words, and the app will display the predicted next word based on the input.

Contributing

Contributions are welcome! Please fork the repository and submit a pull request to contribute to this project.

License

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