ResearchBenchmarksImage-conditioned text-to-speech on CelebV-HQFollow7.9383MCD-DTWFaceTTS7.8991128.1636318.428158.692669May 28, 2026Evaluation ResultsMethodMethodLinksMCD-DTWC-SIMIdentity AccuracyFaceTTS2026.057.93830.904860.32Archonfinetuned=falsefinetuned=false2026.058.9180.911762.23