ResearchBenchmarksImage-conditioned text-to-speech on HDTFFollow7.8128MCD-DTWFaceTTS7.7660248.0817628.39758.713238May 28, 2026Evaluation ResultsMethodMethodLinksMCD-DTWC-SIMIdentity AccuracyFaceTTS2026.057.812888.4457.15Archonfinetuned=falsefinetuned=false2026.058.982290.0259.11