Semantica

This project uses sentence embeddings to visualize and analyze the relationship between different sentences in a 2D or 3D space. It supports two main dimensionality reduction techniques: PCA (Principal Component Analysis) for 3D visualization and t-SNE (t-Distributed Stochastic Neighbor Embedding) for 2D visualization.

Folder Structure

├── README.md
├── requirements.txt
└── src
    ├── app.py
    └── utils.py

Features

• Generate Sentence Embeddings: Using pre-trained models from SentenceTransformers.
• Visualize Embeddings: Visualize the relationship between sentences using 3D PCA or 2D t-SNE.
• Clustering: Optionally apply KMeans clustering to group similar sentences.
• Export Results: Download the projection results as a CSV file.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/sentence-embedding-visualization.git
   cd sentence-embedding-visualization

2. Create a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate 

3. Install the required dependencies:

   pip install -r requirements.txt

Usage

To run the app locally, use the following command:

streamlit run app.py