MahanVeisi8/VAE-MNIST-Variable-Latent-Size-Reconstruction-and-Visualization

Dive into the world of Variational Autoencoders (VAEs) with MNIST! 🎨✨ Explore variable latent sizes (2, 4, 16) to see how they affect reconstruction, latent space visualizations, and performance metrics 📊 (MSE, SSIM, PSNR).

13
/ 100
Experimental

This project helps deep learning practitioners understand how Variational Autoencoders (VAEs) work. It takes images, specifically handwritten digits like those from the MNIST dataset, and shows how a VAE can compress them into a 'latent space' and then reconstruct them. The output includes reconstructed images, visualizations of the latent space, and performance metrics. It's designed for researchers, students, and engineers who are exploring generative models and data compression.

No commits in the last 6 months.

Use this if you want to explore the impact of different latent space sizes on image reconstruction quality and visualize how a VAE organizes data in its hidden dimensions.

Not ideal if you need a production-ready solution for generating complex, high-resolution images or if you're looking for an application-specific generative model beyond basic image data.

deep-learning generative-models data-visualization image-reconstruction machine-learning-research
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 8 / 25
Community 0 / 25

How are scores calculated?

Stars

10

Forks

Language

Jupyter Notebook

License

Last pushed

Jan 22, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/MahanVeisi8/VAE-MNIST-Variable-Latent-Size-Reconstruction-and-Visualization"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.