Skip to content

Latest commit

 

History

History
93 lines (67 loc) · 3.12 KB

README.md

File metadata and controls

93 lines (67 loc) · 3.12 KB

LLM-Generated Text Detector

This project is a web-based application that classifies text as either human-generated or LLM (Large Language Model) generated. The application is built using a Flask backend and a deep learning model implemented with TensorFlow and Keras.

Features

  • Text Classification: The model classifies whether a given text is generated by a human or an LLM.
  • Visualization: The application provides a bar graph displaying the probability distribution of the classification.
  • Word Cloud: A word cloud of the input text is generated and displayed on the frontend. !(./SihLLMEmbedding.jpg) !(./SihLLMEmbedding.jpg) !(./SihLLMEmbedding.jpg)

Model Architecture

The model used for classification is a Sequential neural network with the following layers:

  • Embedding Layer: Converts input text sequences into dense vectors of fixed size.
  • LSTM Layers: Two LSTM (Long Short-Term Memory) layers are used for sequential data processing.
    • LSTM with 128 units (returning sequences).
    • LSTM with 64 units.
  • Dropout Layers: Added to prevent overfitting with a dropout rate of 50%.
  • Dense Layers:
    • Dense layer with 32 units and ReLU activation.
    • Output Dense layer with 1 unit and sigmoid activation.

The model is compiled with the Adam optimizer, binary crossentropy loss, and accuracy as a metric.

Installation

  1. Clone the repository: 
    git clone https://github.com/your-username/LLM-Generated-Text-Detector.git
cd LLM-Generated-Text-Detector

Here is the text converted to markdown:

## Installation

1. **Install dependencies:**

   ```bash
   pip install -r requirements.txt

  1. Download the model:

    Place the pre-trained model (model2.keras) and tokenizer (tokenizer.pkl) in the root directory of the project.

  2. Run the Flask application:

    python app.py

  3. Access the application:

    Open your web browser and go to http://127.0.0.1:5000/.

File Structure

  • app.py: The main Flask application file.
  • preprocess.py: Contains the preprocess function for tokenizing and padding input text.
  • plot_prediction.py: Contains the plot_prediction function for generating and saving the prediction plot and word cloud.
  • templates/: Contains the HTML templates for rendering the web pages.
    • index.html: The home page.
    • predict.html: The prediction result page.
  • static/: Contains static files like CSS, JavaScript, and images.
    • prediction_plot.png: The prediction plot generated during runtime.
    • wordcloud.png: The word cloud image generated from the input text.

Usage

  1. Navigate to the home page and enter the text you want to classify.
  2. Submit the text to get the prediction result.
  3. The result page will display:
    • The classification result (Human-generated or LLM-generated).
    • A bar graph showing the prediction probability.
    • A word cloud generated from the input text.

Dependencies

  • Flask
  • TensorFlow
  • Keras
  • Matplotlib
  • Seaborn
  • WordCloud

License

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