🎧 AudioX-Turbo: Ultra-fast audio generation from any data
The AudioX-Turbo framework has been introduced, which generates high-quality sound in just 4 sampling steps. This is ~25x faster than multi-step counterparts. The model supports generation from text, video, or audio.
🌍 The breakthrough in inference speed allows for the implementation of real-time sound generation for games and video editors.
👤 AI production tools are becoming instantaneous for creating soundtracks and voiceovers.