VAE_Classifier Semi-supervised Learning with Deep Generative Models MNIST analogies(model: M2, latent dim: 50, epochs: 800) 2D Latent generate Image(model: M2) Accuracy N M2 M1+M2 3000 95.89% 95.67% 1000 95.12% 94.35% 600 94.29% 92.99% 100 82.25% 90.66%
