LibreAI / Image / MirrorMetrics
// image · AndyLone22/MirrorMetrics

MirrorMetrics

Visual diagnostics for LoRA training — spot overfitting with face biometrics

💎 Hidden gem★ 62 stars⑂ 8 forksMITupdated 2026-02
▦ visual — this project's README ships no embeddable screenshot,
so here's a placeholder. The guide below is complete regardless.
// project facts

Project facts

MirrorMetrics is an open-source image project by AndyLone22 (AndyLone22/MirrorMetrics on GitHub). It has 62 GitHub stars and 8 forks, is released under the MIT license, and was last updated on GitHub on 21 February 2026.

GitHub stars
62
Forks
8
License
MIT
Last updated on GitHub
21 February 2026
Category
Image
Listing
Hidden gem — a shorter page for a lesser-known project
Official site
Official site
// how it works

How it works

  1. Prepare your LoRA

    Point MirrorMetrics at a trained LoRA and its dataset.

  2. Run diagnostics

    It analyzes samples with InsightFace biometrics.

  3. Read the visuals

    Charts reveal overfitting and dataset problems.

  4. Fix and retrain

    Adjust settings based on what the metrics show.

// features

Why people love it

Overfit detection

Visual signals when a LoRA memorizes.

Dataset quality

Spot bad or inconsistent training images.

Face biometrics

InsightFace-based identity consistency checks.

Training fixes

Guidance on settings to change.

Data visualization

Charts instead of guesswork.

MIT licensed

Free for commercial use with attribution.

Advertisement
// run it

Run it yourself

  1. Step 1 — Clone the repo

    Download the source and move into the project folder.

    git clone https://github.com/AndyLone22/MirrorMetrics.git
    cd MirrorMetrics
  2. Step 2 — Check the README

    Every project documents its own setup; start with install requirements.

    cat README.md | head -60
  3. Step 3 — Run it

    Follow the project's launch command — Docker, pip, npm, or a binary release.

    # see README for the exact launch command
// use cases

What you can build

Advertisement
// faq

Questions, answered

Who maintains MirrorMetrics?

It is an open-source project at AndyLone22/MirrorMetrics with 162 forks. Check the repo's contributors and recent commits to see how active maintenance is.

Is MirrorMetrics free for commercial use?

MIT licensed — commercial use allowed with attribution.

Which models does it evaluate?

It is built for Stable Diffusion LoRAs.

Do I need a GPU?

InsightFace analysis runs best with one — check the README.

Can it fix my LoRA?

It diagnoses problems and suggests setting changes; you retrain.

// keep exploring

Similar tools

Image

kohya_ss

kohya_ss — open-source image project on GitHub.

★ 12.6k starsApache-2.0
View guide →
Image

OneTrainer

OneTrainer is a one-stop solution for all your Diffusion training needs.

★ 3.2k starsAGPL-3.0
View guide →
Image

ai-toolkit

The ultimate training toolkit for finetuning diffusion models

★ 12.2k starsMIT
View guide →
Image

TagPilot

Privacy-first browser tool for tagging LoRA training datasets — no install

★ 40 starsMIT
View guide →
Image

Stable-Diffusion

FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tut…

★ 2.8k starsGPL-3.0
View guide →
// alternatives

Alternatives in Image

The most-starred other Image projects in the LibreAI catalogue — the realistic shortlist if MirrorMetrics is not the right fit.

Get MirrorMetrics

Free and open — grab the source, star the repo, and start creating.