音声
Coqui TTS
Deep learning toolkit for text-to-speech, battle-tested in research and production
★ 46.1k スターMPL-2.0
ガイドを見る →Mozilla TTSは、mozilla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: mozilla/TTS）。GitHubスターは10,174、フォークは1,323で、MPL-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2023年11月9日です。
Mozilla's deep-learning text-to-speech library with Tacotron, Glow-TTS and neural vocoders.
Plan for an older environment: the library supports Python >=3.6 and <3.9, and the project has not been updated since late 2023.
For synthesis only, pip install TTS; to train or modify models, clone the repo and install it editable.
List the bundled pretrained models, then choose a text-to-speech model and a vocoder from the list.
Generate speech from the command line with the tts CLI, or write a config.json for your dataset and run a training script such as train_tacotron.py.
Tacotron, Tacotron2, Glow-TTS and SpeedySpeech implementations, with several attention methods including guided attention and Graves attention.
MelGAN, MultiBand-MelGAN, GAN-TTS, ParallelWaveGAN, WaveGrad and WaveRNN vocoders to turn spectrograms into waveforms.
A GE2E speaker encoder computes speaker embeddings efficiently, enabling multi-speaker models and voice conditioning.
Released models in PyTorch, TensorFlow and TFLite formats, plus a tts command-line tool to list models and synthesize speech in one command.
Fast multi-GPU training via distribute.py, detailed console and TensorBoard logs, fine-tuning with --restore_path, and conversion of PyTorch models to TensorFlow/TFLite.
Tools to analyze and curate TTS datasets, a generic dataloader for custom corpora, a demo server for testing models, and benchmarking notebooks.
Clone Mozilla TTS if you plan to train models or work from source.
git clone https://github.com/mozilla/TTS.git
cd TTSInstall from source in editable mode (for synthesis only, pip install TTS is the documented shortcut - use a Python 3.6-3.8 environment).
pip install -e .List the released models, then run the CLI with a model and vocoder name copied from that list.
tts --list_models
tts --text "Text for TTS" --model_name "///" --vocoder_name "///" --out_path output/ # fill in names from --list_modelsIt was developed under Mozilla; the repository has not seen pushes since 2023-11-09, so today it is effectively community-preserved rather than actively maintained.
Yes. It is released under the Mozilla Public License 2.0, which permits use including in commercial products, with file-level copyleft terms.
No. Everything runs locally - pretrained models download once, and training and synthesis happen on your own machine.
Coqui TTS began as the community continuation of this codebase and is the actively maintained successor with newer models; Mozilla TTS is the original library, best for its classic architectures and recipes.
Create a Python 3.8 environment, run pip install TTS, then tts --list_models and synthesize with one of the released models.
Deep learning toolkit for text-to-speech, battle-tested in research and production
WaveRNN vocoder plus Tacotron TTS in PyTorch: train your own neural speech synthesis on LJSpeech.
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text…
Audiocraft is a library for audio processing and generation with deep learning. It features the state-of-the-art EnCodec audio…
gTTS — Python library and CLI that turns text into spoken mp3 via Google Translate's speech API
Speech-to-text-to-speech for VRChat, VTubers, and streamers.
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 — Mozilla TTS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Mozilla TTS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。
Open research toolkit for text-to-speech, voice conversion, singing voice and text-to-audio generation.
A robust, efficient, low-latency speech-to-text library with advanced voice activity detection, wake word activation and instant transcription.
Silero VAD: pre-trained enterprise-grade Voice Activity Detector
End-to-End Speech Processing Toolkit
Neural building blocks for speaker diarization: speech activity detection, speaker change detection, overlapped speech detection, speaker embedding
Python library for audio and music analysis