Everything is a Tensor

A small visual explainer showing how common data types become PyTorch tensors:

• Tabular data
• Time-series (sequential) data
• Text (sequential) data
• Image (spatial) data

The project is intentionally small and client-side only. It focuses on the basic structural idea: common inputs become arrays of numbers with shapes a model can process.

Live Demo: Everything is a Tensor

Stack

• React
• Vite
• Tailwind CSS
• GitHub Pages

Local Development

npm install
npm run dev

Production Build

npm run build

Deployment

GitHub Actions builds the site and deploys it to GitHub Pages.

Scope

This project deliberately omits modalities like audio and video, for simplicity. It stays on the shortest path to the main idea: everything a model sees is ultimately a tensor.