Lightning-AI / Lightning-AI/torchmetrics

Add Vendi Score and Memorization metrics for generative model evaluation

Open
#3,233 1 comment 1 reaction 1 assignee View on GitHub

@jdipalma2 is already working on this.

Since Sep 9, 2025.

enhancement
Dominant language
Python
Stars
2.5k
Forks
526
Avg merge
6d 11h
Merged PRs (30d)
5

Description

## 🚀 Feature

I would like to add the Vendi Score (https://arxiv.org/abs/2210.02410, https://arxiv.org/abs/2310.12952) and the L2-memorization percentage (https://arxiv.org/abs/2301.13188) for evaluating the performance of generative image models.

### Motivation

These are useful metrics for evaluating the holistic performance of generative models that go beyond the simple FID metric.

### Pitch

I would like to implement these metrics and contribute them to this library.

### Additional context

There are existing implementations of the Vendi Score (https://github.com/vertaix/Vendi-Score) and L2-memorization percentage (https://github.com/layer6ai-labs/dgm-eval) but they aren't incorporated into the torchmetrics API.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.