Thor Schaeff demonstrates audio analysis that labels speakers, languages and timestamps, and explains how structured output can support an application interface.
He shows directed speech generationAudio generation uses AI to create speech, music, sound effects or other audio from instructions and reference inputs. and a real-time multimodal conversation, then demonstrates a voice-driven jukebox that calls a music generation model. He notes where individual demonstrations failed or produced imperfect resultsEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria..
Watch on YouTube
