Unverified50% confidenceFactExact time
SteerTTS-edge's vocoder is a 6.2-million-parameter Vocos-style model distilled from charactr/vocos-mel-24khz, also using int4 QAT.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/22/2026
First Seen
Valid until: 1/20/2027
Sources
Related Entities
Related Claims
UnverifiedSteerTTS edge's vocoder is a 6.2M-parameter Vocos-style model distilled from charactr/vocos-mel-24khz and subjected to int4 QAT.89% similarUnverifiedSteerTTS adopts FastPitch-style explicit variance modeling: phonemes first pass through an encoder, three predictors for duration, pitch and energy each output control quantities, which are then aligned via a length regulator and finally decoded into 100-band log-mel spectra, handed to a frozen Vocos vocoder for waveform synthesis.74% similarUnverifiedThis local voice assistant system is fully self-hosted, with voice detection and recognition running on the CPU, the LLM (Gemma 4 E4B, QAT q4_0 quantization, based on llama.cpp) responsible for generating answers, and the TTS (Qwen3-TTS 1.7B) synthesizing speech sentence by sentence.70% similarUnverifiedSteerTTS-edge's acoustic model is a 9.25-million-parameter attention + bidirectional Mamba-3 architecture, compressed to int4 precision through quantization-aware training (QAT).70% similarUnverified语音编码器 + LLM 解码器的组合是 2023 年以来语音大模型的主流范式，Qwen-Audio、Gemini Audio、Phi-4-audio 等均采用类似设计68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1087444API
curl https://kongchang.com/api/v1/knowledge/claims/1087444MCP
get_claim(id=1087444)