⧢ EVEZ ART GALLERY ⦟⧢

Eigenvalue-conditioned generations from PCA autoencoder trained on 79 artworks by Steven Crawford-Maggard

Model: PCA Autoencoder (16 latent dims, 96.4% variance)
Training: 79 images at 32×32×3, pure NumPy SVD + gradient descent
Conditioning: 6 EVEZ eigenvalues (Φ, η*, r, λ_dom, λ_I-80, r_I-80)
Model size: 397KB
Reconstruction MSE: 0.004057
Top eigenvalue: 114.14 (the dominant component = the primary face)