Predictive Keyword Model

A deep learning-based text prediction tool trained on Sherlock Holmes stories. This project utilizes an LSTM (Long Short-Term Memory) neural network to predict the next word in a sequence, offering a smart "predictive keyboard" experience.

🚀 Features

• Next-Word Prediction: Suggests the most likely next words based on input text.
• Deep Learning Core: Built with PyTorch using Embedding and LSTM layers.
• Interactive Web UI:
  • Streamlit: A modern, clean web interface.
  • Gradio: An alternative quick-prototyping interface.
• Customizable: Easy to retrain on different datasets.

🛠️ Installation

1. Clone the repository:

   git clone https://github.com/sagar-grv/Predictive-word-Model.git
   cd Predictive-word-Model

2. Install dependencies:

   pip install -r requirements.txt

🎮 Usage

Option 1: Streamlit Interface (Recommended)

Run the Streamlit app for a polished user experience.

streamlit run run_model_st.py

Access the app at: http://localhost:8501

Option 2: Gradio Interface

Run the Gradio interface for a simple testing environment.

python run_model.py

Access the app at: http://127.0.0.1:7860

📂 Project Structure

• run_model_st.py: Main Streamlit application file.
• run_model.py: Gradio application file.
• predictive_keyword_model.ipynb: Original research and training notebook.
• predictive_keyboard_model.pth: Pre-trained model weights.
• requirements.txt: Python package dependencies.
• sherlock-holm.es_stories_plain-text_advs.txt: Training dataset.

🧠 Model Architecture

• Embedding Layer: Converts words into dense vectors.
• LSTM Layer: Captures sequential dependencies in the text.
• Fully Connected Layer: Maps LSTM output to vocabulary size for prediction.

---

Created by Sagar grv