Facetron: A Multi-speaker Face-to-Speech Model based on Cross-modal Latent Representations

Open in new window