mingukkang/elatentlpips

Author's Implementation for E-LatentLPIPS

41
/ 100
Emerging

This project helps researchers and developers working with image generation models quickly measure how perceptually similar two images are, directly in their compressed 'latent' form. It takes two compressed image representations from a Latent Diffusion Model and outputs a single number indicating their perceptual distance. It's designed for machine learning researchers and engineers who need to evaluate image generation quality or perform regression tasks without resource-intensive image decoding.

179 stars. No commits in the last 6 months. Available on PyPI.

Use this if you need a faster, memory-efficient way to calculate perceptual similarity between images generated by Latent Diffusion Models like Stable Diffusion or FLUX, especially during training or evaluation.

Not ideal if your workflow requires human-interpretable pixel-level comparisons or if you are not working with latent representations from diffusion models.

image-generation perceptual-similarity latent-space-analysis computer-vision deep-learning-evaluation
Stale 6m
Maintenance 0 / 25
Adoption 10 / 25
Maturity 25 / 25
Community 6 / 25

How are scores calculated?

Stars

179

Forks

4

Language

Python

License

Last pushed

Nov 05, 2024

Commits (30d)

0

Dependencies

7

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/mingukkang/elatentlpips"

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