FID EXPLORER

Visualizing the Fréchet Inception Distance. We evaluate Generative AI not by comparing pixels, but by extracting features using an Inception Network and comparing their statistical distributions in a latent space.

Total FID Score
0.00
Evaluating...

AI Model Tuning

Feature Averages ($\mu_g$)

Shift X 0
Shift Y 0

Feature Variance ($\Sigma_g$)

Diversity X 0
Diversity Y 0

Training Scenarios

Real Images (Baseline)
AI Generated (Current)

The FID Formula

$$ d^2 = ||\mu_r - \mu_g||^2 + Tr(\Sigma_r + \Sigma_g - 2\sqrt{\Sigma_r\Sigma_g}) $$
Mean Penalty 0.0

Euclidean distance between centers $||\mu_r - \mu_g||^2$. Penalizes AI for generating the wrong *kind* of features.

Covariance Penalty 0.0

Difference in distribution shapes (Trace term). Penalizes AI for lacking diversity or having the wrong variation.

Why is it called "Inception"?

We don't calculate this on raw pixels. We pass images through a pre-trained image classifier called InceptionV3. We chop off the final classification layer and look at the 2,048-dimensional feature vector it produces. This visualizer is a 2D simplification of that 2,048D space.