ChipsXDAabout 7 hours ago
Voice cloned in seconds with no GPU needed
1 min read
Kyutai released Pocket TTS, a voice cloning model that runs on a CPU. A five-second audio sample produced a convincing clone in under 15 minutes using a Ryzen 5 7600 CPU with no GPU. The model generated audio in 46 milliseconds, raising concerns about banking voice verification security.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
