gabriele-mastrapasqua/qwen3-tts

Real-time text-to-speech with Qwen3-TTS

Rust implementation of Qwen3-TTS speech synthesis

Ultrafast Qwen3-TTS: sub-50 ms time-to-first-audio at 10 requests per second.

Local AI text-to-speech with voice cloning and voice design, powered by GGML. C++17 port of Qwen3-TTS (QwenLM/Qwen3-TTS). 10 languages, 24 kHz mono output, runs on CPU, CUDA, ROCm, Metal, Vulkan.

基于 Qwen3 TTS 的 WebUI

ComfyUI custom nodes for speech, voice cloning, and voice design based on Qwen3-TTS models

Triton kernel fusion for Qwen3-TTS 1.7B inference acceleration — RMSNorm, SwiGLU, M-RoPE, Norm+Residual

Qwen3-TTS is an open-source series of TTS models developed by the Qwen team at Alibaba Cloud, supporting stable, expressive, and streaming speech generation, free-form voice design, and vivid voice cloning.

最极速的Qwen3-TTS推理方案。将 Qwen3-TTS 的 LLM 部分导出为 GGUF，用 llama.cpp 进行加速推理。后者支持 Vulkan 和 Cuda 加速。