音声
audio.cpp
A pure C++ ggml inference engine for audio models — TTS, STT, VAD, voice conversion and music.
★ 3.4k スターNOASSERTION
ガイドを見る →GPAは、AutoArkが公開しているオープンソースの音声プロジェクトです（GitHub: AutoArk/GPA）。GitHubスターは3,109、フォークは206で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2026年5月25日です。
GPA — one tiny model that does ASR, TTS and voice conversion (General Purpose Audio).
Decide the first concrete outcome you want from it, so you can pick the right example, template or mode instead of exploring everything at once.
Clone the repository (or install the published package) and skim its README quick start for the requirements for your platform.
Run the documented install step and set any required configuration — such as provider keys, data sources or local services — before the first run.
Execute the smallest documented example end to end, confirm the output is what you expected, then adapt it to your real data or workflow.
ASR, TTS and voice conversion in a single tiny model instead of three separate systems.
Native inference and training pipelines with published TTS/ASR evaluation tables.
GPA-TTS ships INT8/FP16/FP32 and INT4 ONNX variants for constrained hardware.
CLI, API and UI around the v1.5 ONNX runtime for easy integration.
GPA-0.3B-preview and GPA-v1.5 checkpoints are released, with a public roadmap (deployment recipes and a native VC path are listed as next steps).
An LLM-based design for both audio generation and understanding.
Get the source so you can inspect, configure and run it locally.
git clone https://github.com/AutoArk/GPA.git
cd GPAClone the repo and follow the GPA-v1.5 (GPA_1.5/) or GPA-TTS (GPA_TTS/) folder's own README for environment setup, checkpoints and inference commands.
git clone https://github.com/AutoArk/GPA.git
cd GPA # see GPA_1.5/README.md or GPA_TTS/README.md for the exact commandsUse the v1.5 ONNX Runtime CLI/API/UI or the GPA-TTS runtime as documented in their sub-READMEs.
# see README for the exact command in GPA_1.5/ or GPA_TTS/It is maintained by AutoArk on GitHub (AutoArk/GPA), where the source, issues and releases live.
Yes — it is free and open source under Apache-2.0, which permits use, modification and commercial redistribution, with licence/notice retention.
No — GPA runs locally from released checkpoints; there is no hosted API key involved.
Compared with separate specialist models for ASR, TTS and VC, GPA's distinguishing focus is: gPA — one tiny model that does ASR, TTS and voice conversion (General Purpose Audio). Pick it when that focus matches your need; the alternative may suit you better if its broader or different scope is the priority.
Clone https://github.com/AutoArk/GPA (or install its published package), follow the README quick start to install and configure it, then run the smallest documented example — the Run section on this page mirrors those steps.
A pure C++ ggml inference engine for audio models — TTS, STT, VAD, voice conversion and music.
MARS5 speech model from CAMB.AI — clone a voice from 5 seconds of audio with striking prosody
A tiny conversational TTS model with cloning, streaming, and CPU speed.
A mental-model-first guide from waveforms to voice systems.
One Gradio + React WebUI for dozens of TTS, voice-cloning and music models — Bark to Kokoro.
Easily train a good VC model with voice data <= 10 mins!
LibreAIカタログに掲載された音声プロジェクトの中で最もスターの多いもの — GPAが合わなかった場合に検討したい、現実的な候補リストです。
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Robust Speech Recognition via Large-Scale Weak Supervision
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
Port of OpenAI's Whisper model in C/C++
カタログの中でGPAとスター数が近い音声プロジェクト — 前後の棚もそのまま見て回れます。
One Gradio + React WebUI for dozens of TTS, voice-cloning and music models — Bark to Kokoro.
Unofficial PyTorch implementation of VALL-E, the neural codec language model for zero-shot TTS
A pure C++ ggml inference engine for audio models — TTS, STT, VAD, voice conversion and music.
MARS5 speech model from CAMB.AI — clone a voice from 5 seconds of audio with striking prosody
A simple, high-quality RVC voice conversion tool focused on ease of use and performance.
Vietnamese TTS with instant voice cloning — on-device, real-time CPU inference, 48 kHz audio