KittenML has released Kitten-TTS 2, an open-source text-to-speech model that utilizes a novel 1.58-bit ternary weight system. This allows for real-time performance on CPUs without requiring a GPU, making it accessible for users with limited hardware. The model supports zero-shot voice cloning in over 20 languages, multilingual capabilities, and allows for expressive control over speech delivery, offering an alternative to commercial TTS services. AI
IMPACT Provides a highly efficient, open-source TTS solution for users with limited hardware, enabling local voice cloning and multilingual generation.
RANK_REASON Release of an open-source model with novel technical specifications. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →