A fast, local neural text to speech system that sounds great and is optimized for the Raspberry Pi 4.
Piper is used in a [variety of projects](#people-using-piper).
## 目次
- [追加機能](#追加機能)
- [関連記事](#関連記事)
- [プラットフォームサポート](#プラットフォームサポート)
- [対応プラットフォーム](#対応プラットフォーム)
- [⚠️ 重要: macOSユーザーへのお知らせ](#️-重要-macosユーザーへのお知らせ)
- [Voices](#voices)
- [Installation](#installation)
- [Usage](#usage)
- [Streaming Audio](#streaming-audio)
- [JSON Input](#json-input)
- [People using Piper](#people-using-piper)
- [Training](#training)
- [Running in Python](#running-in-python)
## 追加機能
* 日本語の事前学習及び追加学習/推論対応（OpenJTalk統合）
* 詳細な使用方法は[日本語音声合成ガイド](JAPANESE_USAGE.md)を参照
* **Windows対応**: [Windowsセットアップガイド](docs/windows-setup.md)を参照
* **API ドキュメント**: [OpenJTalk API リファレンス](docs/openjtalk-api.md)を参照
* PUA音素マッピングによる日本語TTS精度向上 - [技術詳細](PHONEME_MAPPING.md)を参照
* **自動ダウンロード機能**: 初回実行時に必要な辞書とHTSボイスファイルを自動ダウンロード
* 環境変数（オプション）：
- `OPENJTALK_DICTIONARY_DIR`: OpenJTalk辞書へのパス（未設定時は自動ダウンロード）
- `OPENJTALK_VOICE`: HTSボイスモデル（.htsvoice）へのパス（未設定時は自動ダウンロード）
- `PIPER_AUTO_DOWNLOAD_DICT`: `0`に設定すると自動ダウンロードを無効化
- `PIPER_OFFLINE_MODE`: `1`に設定するとオフラインモード（ネットワーク接続不要）
* 既存の日本語モデルは**再学習不要** - 設定ファイルの更新のみで対応可能
* GitHub Actionsによる自動ビルドとバイナリー配布（詳細は[プラットフォームサポート](#プラットフォームサポート)を参照）
* 前処理済み .pt ファイルが破損していても学習時に自動スキップして継続できるように改善
* DataLoader に `pin_memory=True` を設定し GPU 転送を最適化
* `preprocess.py` に `--timeout-seconds` を追加し、ハングする発話を自動タイムアウト/スキップ
* `piper_train` に `--num-workers` を追加し、DataLoader のワーカー数をコマンドラインから指定可能に
* `piper_train` に `--save-top-k` を追加し、チェックポイント保存個数をコマンドラインから指定可能に
* PyPI パッケージ `piper-tts-plus` として公開し、`pip install` で簡単インストール可能に
* 多言語TTSテストインフラストラクチャーを追加し、CI/CDで自動テスト実行 - [詳細](docs/MULTILINGUAL_TESTING.md)
* OpenJTalk辞書とHTSボイスモデルの自動ダウンロード機能を追加し、日本語TTSのセットアップを簡略化
* マルチGPU学習対応（PyTorch Lightning 2.x）
* DDP (Distributed Data Parallel) 戦略による複数GPU並列学習
* 学習率の自動スケーリング機能（`--auto_lr_scaling`）
* 使用例：
```bash
python -m piper_train \
--dataset-dir /path/to/dataset \
--batch-size 16 \
--devices 2 \
--strategy ddp \
--base_lr 2e-4
# 注: --auto_lr_scaling はデフォルトで有効
# 無効にする場合は --disable_auto_lr_scaling を使用
```
* チェックポイント管理機能の強化
* `--resume_from_checkpoint` でチェックポイントからの学習再開
* `--resume_from_single_speaker_checkpoint` でシングルスピーカーモデルからマルチスピーカーへの変換
* GPU推論サポート（C++バイナリ）
* `--use-cuda` オプションでONNX Runtime CUDAプロバイダーを有効化
* 学習時の高度なオプション
* `--gradient_clip_val` - 勾配クリッピング
* `--accumulate_grad_batches` - 勾配累積によるバッチサイズ仮想拡張
* `--precision` - Mixed Precision Training対応（16-mixed等）
* `--detect_anomaly` - 学習時の異常検出機能
* 音声評価ツール（`scripts/evaluation/`）
* MCD (Mel-Cepstral Distortion) 評価
* PESQ (Perceptual Evaluation of Speech Quality) 評価
* UTMOS評価
## 関連記事
* [LJSpeechを使って英語のpiperの事前学習モデルを作成する](https://ayousanz.hatenadiary.jp/entry/2025/05/26/230341)
* [jvs音声データセットを使ったpiper日本語モデルの作成](https://ayousanz.hatenadiary.jp/entry/2025/06/05/093217)
* [piperモデルからつくよみちゃんデータセットを使って追加学習を行う](https://ayousanz.hatenadiary.jp/entry/2025/06/07/074232)
``` sh
echo 'Welcome to the world of speech synthesis!' | \
./piper --model en_US-lessac-medium.onnx --output_file welcome.wav
```
[Listen to voice samples](https://rhasspy.github.io/piper-samples) and check out a [video tutorial by Thorsten Müller](https://youtu.be/rjq5eZoWWSo)
Voices are trained with [VITS](https://github.com/jaywalnut310/vits/) and exported to the [onnxruntime](https://onnxruntime.ai/).
[](https://www.openhomefoundation.org/)
## プラットフォームサポート
### 対応プラットフォーム
| プラットフォーム | アーキテクチャ | OpenJTalk対応 | 備考 |
|-----------------|---------------|--------------|------|
| Linux | x86_64 (amd64) | ✅ | フルサポート |
| Linux | ARM64 | ✅ | フルサポート (CMakeビルド使用) |
| macOS | **ARM64 (Apple Silicon)のみ** | ✅ | M1/M2/M3以降のMac専用 |
| Windows | x64 | ✅ | フルサポート |
### ⚠️ 重要: macOSユーザーへのお知らせ
**2024年より、macOSではApple Silicon (M1/M2/M3以降) のみをサポートしています。**
#### Intel Macをお使いの方へ
Intel Mac (x86_64) のサポートは終了しました。以下の代替方法をご利用ください：
1. **Dockerを使用（推奨）**
```bash
# Dockerイメージをプル
docker pull ghcr.io/ayutaz/piper-plus:latest
# 実行例
docker run --rm -v $(pwd):/data ghcr.io/ayutaz/piper-plus:latest \
echo "Hello from Docker" | piper --model /data/model.onnx --output_file /data/output.wav
```
2. **ソースからビルド**
```bash
# 依存関係をインストール
brew install cmake onnxruntime
# ビルド
git clone https://github.com/ayutaz/piper-plus.git
cd piper-plus
mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j$(sysctl -n hw.ncpu)
```
3. **仮想マシンでLinux版を使用**
- UTM、Parallels Desktop、VMware Fusionなどを使用
#### Apple Siliconユーザーの方へ
通常通りダウンロードしてご利用いただけます。初回実行時のセキュリティ警告については、以下をご参照ください。
##### macOSセキュリティ警告の対処
ダウンロードしたバイナリを初めて実行する際、macOSのセキュリティ機能により警告が表示される場合があります。以下のコマンドで検疫属性を削除してください：
```bash
# ダウンロードしたファイルを展開後
xattr -cr piper/
# または特定のバイナリのみ
xattr -cr piper/bin/piper
xattr -cr piper/bin/open_jtalk # 日本語TTSを使用する場合
```
これにより、Gatekeeperの警告なしに実行できるようになります。
## Voices
Our goal is to support Home Assistant and the [Year of Voice](https://www.home-assistant.io/blog/2022/12/20/year-of-voice/).
[Download voices](VOICES.md) for the supported languages:
* العربية, Jordan (Arabic, ar_JO)
* Català, Spain (Catalan, ca_ES)
* Čeština, Czech Republic (Czech, cs_CZ)
* Cymraeg, Great Britain (Welsh, cy_GB)
* Dansk, Denmark (Danish, da_DK)
* Deutsch, Germany (German, de_DE)
* Ελληνικά, Greece (Greek, el_GR)
* English, Great Britain (English, en_GB)
* English, United States (English, en_US)
* Español, Argentina (Spanish, es_AR)
* Español, Spain (Spanish, es_ES)
* Español, Mexico (Spanish, es_MX)
* فارسی, Iran (Farsi, fa_IR)
* Suomi, Finland (Finnish, fi_FI)
* Français, France (French, fr_FR)
* Magyar, Hungary (Hungarian, hu_HU)
* íslenska, Iceland (Icelandic, is_IS)
* Italiano, Italy (Italian, it_IT)
* ქართული ენა, Georgia (Georgian, ka_GE)
* қазақша, Kazakhstan (Kazakh, kk_KZ)
* Lëtzebuergesch, Luxembourg (Luxembourgish, lb_LU)
* Latviešu, Latvia (Latvian, lv_LV)
* മലയാളം, India (Malayalam, ml_IN)
* हिंदी, India (Hindi, hi_IN)
* नेपाली, Nepal (Nepali, ne_NP)
* Nederlands, Belgium (Dutch, nl_BE)
* Nederlands, Netherlands (Dutch, nl_NL)
* Norsk, Norway (Norwegian, no_NO)
* Polski, Poland (Polish, pl_PL)
* Português, Brazil (Portuguese, pt_BR)
* Português, Portugal (Portuguese, pt_PT)
* Română, Romania (Romanian, ro_RO)
* Русский, Russia (Russian, ru_RU)
* Slovenčina, Slovakia (Slovak, sk_SK)
* Slovenščina, Slovenia (Slovenian, sl_SI)
* srpski, Serbia (Serbian, sr_RS)
* Svenska, Sweden (Swedish, sv_SE)
* Kiswahili, Democratic Republic of the Congo (Swahili, sw_CD)
* Türkçe, Turkey (Turkish, tr_TR)
* украї́нська мо́ва, Ukraine (Ukrainian, uk_UA)
* Tiếng Việt, Vietnam (Vietnamese, vi_VN)
* 简体中文, China (Chinese, zh_CN)
You will need two files per voice:
1. A `.onnx` model file, such as [`en_US-lessac-medium.onnx`](https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/en/en_US/lessac/medium/en_US-lessac-medium.onnx)
2. A `.onnx.json` config file, such as [`en_US-lessac-medium.onnx.json`](https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/en/en_US/lessac/medium/en_US-lessac-medium.onnx.json)
The `MODEL_CARD` file for each voice contains important licensing information. Piper is intended for text to speech research, and does not impose any additional restrictions on voice models. Some voices may have restrictive licenses, however, so please review them carefully!
## Installation
You can [run Piper with Python](#running-in-python) or download a binary release:
* [amd64](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_amd64.tar.gz) (64-bit desktop Linux)
* [arm64](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_arm64.tar.gz) (64-bit Raspberry Pi 4)
* [armv7](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_armv7.tar.gz) (32-bit Raspberry Pi 3/4)
### Building from Source
If you want to build from source, see the [Makefile](Makefile) and [C++ source](src/cpp).
#### Prerequisites
* C++ compiler with C++17 support
* CMake 3.13 or later
* Git
#### Build Steps
1. Clone the repository:
```bash
git clone https://github.com/rhasspy/piper.git
cd piper
```
2. Create build directory:
```bash
mkdir build
cd build
```
3. Configure and build:
```bash
cmake ..
cmake --build . --config Release
```
#### Platform-specific Notes
**Linux**: You must download and extract [piper-phonemize](https://github.com/rhasspy/piper-phonemize) to `lib/Linux-$(uname -m)/piper_phonemize` before building.
For example, `lib/Linux-x86_64/piper_phonemize/lib/libpiper_phonemize.so` should exist for AMD/Intel machines.
**Windows**: See the [Windows Setup Guide](docs/windows-setup.md) for detailed instructions.
**macOS**: The build process will automatically download required dependencies.
## Usage
1. [Download a voice](#voices) and extract the `.onnx` and `.onnx.json` files
2. Run the `piper` binary with text on standard input, `--model /path/to/your-voice.onnx`, and `--output_file output.wav`
For example:
``` sh
echo 'Welcome to the world of speech synthesis!' | \
./piper --model en_US-lessac-medium.onnx --output_file welcome.wav
```
For multi-speaker models, use `--speaker <number>` to change speakers (default: 0).
### Additional Options
* `--use-cuda` - Enable GPU acceleration with CUDA
* `--quiet` / `-q` - Disable logging output
* `--phoneme-silence <phoneme> <seconds>` - Set silence duration for specific phonemes
* `--length-scale <value>` - Adjust speech speed (default: 1.0, smaller = faster)
* `--noise-scale <value>` - Control audio variation (default: 0.667)
* `--noise-w <value>` - Control phoneme duration variation (default: 0.8)
* `--sentence-silence <seconds>` - Silence between sentences (default: 0.2)
See `piper --help` for more options.
### Streaming Audio
Piper can stream raw audio to stdout as its produced:
``` sh
echo 'This sentence is spoken first. This sentence is synthesized while the first sentence is spoken.' | \
./piper --model en_US-lessac-medium.onnx --output-raw | \
aplay -r 22050 -f S16_LE -t raw -
```
This is **raw** audio and not a WAV file, so make sure your audio player is set to play 16-bit mono PCM samples at the correct sample rate for the voice.
### JSON Input
The `piper` executable can accept JSON input when using the `--json-input` flag. Each line of input must be a JSON object with `text` field. For example:
``` json
{ "text": "First sentence to speak." }
{ "text": "Second sentence to speak." }
```
Optional fields include:
* `speaker` - string
* Name of the speaker to use from `speaker_id_map` in config (multi-speaker voices only)
* `speaker_id` - number
* Id of speaker to use from 0 to number of speakers - 1 (multi-speaker voices only, overrides "speaker")
* `output_file` - string
* Path to output WAV file
The following example writes two sentences with different speakers to different files:
``` json
{ "text": "First speaker.", "speaker_id": 0, "output_file": "/tmp/speaker_0.wav" }
{ "text": "Second speaker.", "speaker_id": 1, "output_file": "/tmp/speaker_1.wav" }
```
## People using Piper
Piper has been used in the following projects/papers:
* [Home Assistant](https://github.com/home-assistant/addons/blob/master/piper/README.md)
* [Rhasspy 3](https://github.com/rhasspy/rhasspy3/)
* [NVDA - NonVisual Desktop Access](https://www.nvaccess.org/post/in-process-8th-may-2023/#voices)
* [Image Captioning for the Visually Impaired and Blind: A Recipe for Low-Resource Languages](https://www.techrxiv.org/articles/preprint/Image_Captioning_for_the_Visually_Impaired_and_Blind_A_Recipe_for_Low-Resource_Languages/22133894)
* [Open Voice Operating System](https://github.com/OpenVoiceOS/ovos-tts-plugin-piper)
* [JetsonGPT](https://github.com/shahizat/jetsonGPT)
* [LocalAI](https://github.com/go-skynet/LocalAI)
* [Lernstick EDU / EXAM: reading clipboard content aloud with language detection](https://lernstick.ch/)
* [Natural Speech - A plugin for Runelite, an OSRS Client](https://github.com/phyce/rl-natural-speech)
* [mintPiper](https://github.com/evuraan/mintPiper)
* [Vim-Piper](https://github.com/wolandark/vim-piper)
## Training
See the [training guide](TRAINING.md) and the [source code](src/python).
Pretrained checkpoints are available on [Hugging Face](https://huggingface.co/datasets/rhasspy/piper-checkpoints/tree/main)
## Running in Python
See [src/python_run](src/python_run)
Install with `pip`:
``` sh
# 基本機能のみ
pip install piper-tts-plus
# GPU 版 (CUDA 環境がある場合)
pip install "piper-tts-plus[gpu]"
# HTTP サーバー機能を含む場合
pip install "piper-tts-plus[http]"
# GPU + HTTP
pip install "piper-tts-plus[gpu,http]"
```
This will automatically download [voice files](https://huggingface.co/rhasspy/piper-voices/tree/v1.0.0) the first time they're used. Use `--data-dir` and `--download-dir` to adjust where voices are found/downloaded.
If you'd like to use a GPU, install the `onnxruntime-gpu` package:
``` sh
.venv/bin/pip3 install onnxruntime-gpu
```
and then run `piper` with the `--cuda` argument. You will need to have a functioning CUDA environment, such as what's available in [NVIDIA's PyTorch containers](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch).
Raw data
{
"_id": null,
"home_page": "https://github.com/ayutaz/piper-plus",
"name": "piper-tts-plus",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.11",
"maintainer_email": null,
"keywords": "piper-plus tts text-to-speech japanese openjtalk neural-tts vits multi-gpu pytorch-lightning onnx raspberry-pi",
"author": "yousan",
"author_email": "rabbitcats77@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/46/4f/f4f2b6dd28abcf699f2145d7a958afc361cd4f9ce678bda3a3e8b659f6e2/piper_tts_plus-1.3.0.tar.gz",
"platform": null,
"description": "\n\nA fast, local neural text to speech system that sounds great and is optimized for the Raspberry Pi 4.\nPiper is used in a [variety of projects](#people-using-piper).\n\n## \u76ee\u6b21\n- [\u8ffd\u52a0\u6a5f\u80fd](#\u8ffd\u52a0\u6a5f\u80fd)\n- [\u95a2\u9023\u8a18\u4e8b](#\u95a2\u9023\u8a18\u4e8b)\n- [\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u30b5\u30dd\u30fc\u30c8](#\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u30b5\u30dd\u30fc\u30c8)\n - [\u5bfe\u5fdc\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0](#\u5bfe\u5fdc\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0)\n - [\u26a0\ufe0f \u91cd\u8981: macOS\u30e6\u30fc\u30b6\u30fc\u3078\u306e\u304a\u77e5\u3089\u305b](#\ufe0f-\u91cd\u8981-macos\u30e6\u30fc\u30b6\u30fc\u3078\u306e\u304a\u77e5\u3089\u305b)\n- [Voices](#voices)\n- [Installation](#installation)\n- [Usage](#usage)\n - [Streaming Audio](#streaming-audio)\n - [JSON Input](#json-input)\n- [People using Piper](#people-using-piper)\n- [Training](#training)\n- [Running in Python](#running-in-python)\n\n## \u8ffd\u52a0\u6a5f\u80fd\n* \u65e5\u672c\u8a9e\u306e\u4e8b\u524d\u5b66\u7fd2\u53ca\u3073\u8ffd\u52a0\u5b66\u7fd2/\u63a8\u8ad6\u5bfe\u5fdc\uff08OpenJTalk\u7d71\u5408\uff09\n * \u8a73\u7d30\u306a\u4f7f\u7528\u65b9\u6cd5\u306f[\u65e5\u672c\u8a9e\u97f3\u58f0\u5408\u6210\u30ac\u30a4\u30c9](JAPANESE_USAGE.md)\u3092\u53c2\u7167\n * **Windows\u5bfe\u5fdc**: [Windows\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u30ac\u30a4\u30c9](docs/windows-setup.md)\u3092\u53c2\u7167\n * **API \u30c9\u30ad\u30e5\u30e1\u30f3\u30c8**: [OpenJTalk API \u30ea\u30d5\u30a1\u30ec\u30f3\u30b9](docs/openjtalk-api.md)\u3092\u53c2\u7167\n * PUA\u97f3\u7d20\u30de\u30c3\u30d4\u30f3\u30b0\u306b\u3088\u308b\u65e5\u672c\u8a9eTTS\u7cbe\u5ea6\u5411\u4e0a - [\u6280\u8853\u8a73\u7d30](PHONEME_MAPPING.md)\u3092\u53c2\u7167\n * **\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u6a5f\u80fd**: \u521d\u56de\u5b9f\u884c\u6642\u306b\u5fc5\u8981\u306a\u8f9e\u66f8\u3068HTS\u30dc\u30a4\u30b9\u30d5\u30a1\u30a4\u30eb\u3092\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\n * \u74b0\u5883\u5909\u6570\uff08\u30aa\u30d7\u30b7\u30e7\u30f3\uff09\uff1a\n - `OPENJTALK_DICTIONARY_DIR`: OpenJTalk\u8f9e\u66f8\u3078\u306e\u30d1\u30b9\uff08\u672a\u8a2d\u5b9a\u6642\u306f\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\uff09\n - `OPENJTALK_VOICE`: HTS\u30dc\u30a4\u30b9\u30e2\u30c7\u30eb\uff08.htsvoice\uff09\u3078\u306e\u30d1\u30b9\uff08\u672a\u8a2d\u5b9a\u6642\u306f\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\uff09\n - `PIPER_AUTO_DOWNLOAD_DICT`: `0`\u306b\u8a2d\u5b9a\u3059\u308b\u3068\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3092\u7121\u52b9\u5316\n - `PIPER_OFFLINE_MODE`: `1`\u306b\u8a2d\u5b9a\u3059\u308b\u3068\u30aa\u30d5\u30e9\u30a4\u30f3\u30e2\u30fc\u30c9\uff08\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u63a5\u7d9a\u4e0d\u8981\uff09\n * \u65e2\u5b58\u306e\u65e5\u672c\u8a9e\u30e2\u30c7\u30eb\u306f**\u518d\u5b66\u7fd2\u4e0d\u8981** - \u8a2d\u5b9a\u30d5\u30a1\u30a4\u30eb\u306e\u66f4\u65b0\u306e\u307f\u3067\u5bfe\u5fdc\u53ef\u80fd\n* GitHub Actions\u306b\u3088\u308b\u81ea\u52d5\u30d3\u30eb\u30c9\u3068\u30d0\u30a4\u30ca\u30ea\u30fc\u914d\u5e03\uff08\u8a73\u7d30\u306f[\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u30b5\u30dd\u30fc\u30c8](#\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u30b5\u30dd\u30fc\u30c8)\u3092\u53c2\u7167\uff09\n* \u524d\u51e6\u7406\u6e08\u307f .pt \u30d5\u30a1\u30a4\u30eb\u304c\u7834\u640d\u3057\u3066\u3044\u3066\u3082\u5b66\u7fd2\u6642\u306b\u81ea\u52d5\u30b9\u30ad\u30c3\u30d7\u3057\u3066\u7d99\u7d9a\u3067\u304d\u308b\u3088\u3046\u306b\u6539\u5584\n* DataLoader \u306b `pin_memory=True` \u3092\u8a2d\u5b9a\u3057 GPU \u8ee2\u9001\u3092\u6700\u9069\u5316\n* `preprocess.py` \u306b `--timeout-seconds` \u3092\u8ffd\u52a0\u3057\u3001\u30cf\u30f3\u30b0\u3059\u308b\u767a\u8a71\u3092\u81ea\u52d5\u30bf\u30a4\u30e0\u30a2\u30a6\u30c8/\u30b9\u30ad\u30c3\u30d7\n* `piper_train` \u306b `--num-workers` \u3092\u8ffd\u52a0\u3057\u3001DataLoader \u306e\u30ef\u30fc\u30ab\u30fc\u6570\u3092\u30b3\u30de\u30f3\u30c9\u30e9\u30a4\u30f3\u304b\u3089\u6307\u5b9a\u53ef\u80fd\u306b\n* `piper_train` \u306b `--save-top-k` \u3092\u8ffd\u52a0\u3057\u3001\u30c1\u30a7\u30c3\u30af\u30dd\u30a4\u30f3\u30c8\u4fdd\u5b58\u500b\u6570\u3092\u30b3\u30de\u30f3\u30c9\u30e9\u30a4\u30f3\u304b\u3089\u6307\u5b9a\u53ef\u80fd\u306b\n* PyPI \u30d1\u30c3\u30b1\u30fc\u30b8 `piper-tts-plus` \u3068\u3057\u3066\u516c\u958b\u3057\u3001`pip install` \u3067\u7c21\u5358\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u53ef\u80fd\u306b\n* \u591a\u8a00\u8a9eTTS\u30c6\u30b9\u30c8\u30a4\u30f3\u30d5\u30e9\u30b9\u30c8\u30e9\u30af\u30c1\u30e3\u30fc\u3092\u8ffd\u52a0\u3057\u3001CI/CD\u3067\u81ea\u52d5\u30c6\u30b9\u30c8\u5b9f\u884c - [\u8a73\u7d30](docs/MULTILINGUAL_TESTING.md)\n* OpenJTalk\u8f9e\u66f8\u3068HTS\u30dc\u30a4\u30b9\u30e2\u30c7\u30eb\u306e\u81ea\u52d5\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u6a5f\u80fd\u3092\u8ffd\u52a0\u3057\u3001\u65e5\u672c\u8a9eTTS\u306e\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u3092\u7c21\u7565\u5316\n* \u30de\u30eb\u30c1GPU\u5b66\u7fd2\u5bfe\u5fdc\uff08PyTorch Lightning 2.x\uff09\n * DDP (Distributed Data Parallel) \u6226\u7565\u306b\u3088\u308b\u8907\u6570GPU\u4e26\u5217\u5b66\u7fd2\n * \u5b66\u7fd2\u7387\u306e\u81ea\u52d5\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0\u6a5f\u80fd\uff08`--auto_lr_scaling`\uff09\n * \u4f7f\u7528\u4f8b\uff1a\n ```bash\n python -m piper_train \\\n --dataset-dir /path/to/dataset \\\n --batch-size 16 \\\n --devices 2 \\\n --strategy ddp \\\n --base_lr 2e-4\n # \u6ce8: --auto_lr_scaling \u306f\u30c7\u30d5\u30a9\u30eb\u30c8\u3067\u6709\u52b9\n # \u7121\u52b9\u306b\u3059\u308b\u5834\u5408\u306f --disable_auto_lr_scaling \u3092\u4f7f\u7528\n ```\n* \u30c1\u30a7\u30c3\u30af\u30dd\u30a4\u30f3\u30c8\u7ba1\u7406\u6a5f\u80fd\u306e\u5f37\u5316\n * `--resume_from_checkpoint` \u3067\u30c1\u30a7\u30c3\u30af\u30dd\u30a4\u30f3\u30c8\u304b\u3089\u306e\u5b66\u7fd2\u518d\u958b\n * `--resume_from_single_speaker_checkpoint` \u3067\u30b7\u30f3\u30b0\u30eb\u30b9\u30d4\u30fc\u30ab\u30fc\u30e2\u30c7\u30eb\u304b\u3089\u30de\u30eb\u30c1\u30b9\u30d4\u30fc\u30ab\u30fc\u3078\u306e\u5909\u63db\n* GPU\u63a8\u8ad6\u30b5\u30dd\u30fc\u30c8\uff08C++\u30d0\u30a4\u30ca\u30ea\uff09\n * `--use-cuda` \u30aa\u30d7\u30b7\u30e7\u30f3\u3067ONNX Runtime CUDA\u30d7\u30ed\u30d0\u30a4\u30c0\u30fc\u3092\u6709\u52b9\u5316\n* \u5b66\u7fd2\u6642\u306e\u9ad8\u5ea6\u306a\u30aa\u30d7\u30b7\u30e7\u30f3\n * `--gradient_clip_val` - \u52fe\u914d\u30af\u30ea\u30c3\u30d4\u30f3\u30b0\n * `--accumulate_grad_batches` - \u52fe\u914d\u7d2f\u7a4d\u306b\u3088\u308b\u30d0\u30c3\u30c1\u30b5\u30a4\u30ba\u4eee\u60f3\u62e1\u5f35\n * `--precision` - Mixed Precision Training\u5bfe\u5fdc\uff0816-mixed\u7b49\uff09\n * `--detect_anomaly` - \u5b66\u7fd2\u6642\u306e\u7570\u5e38\u691c\u51fa\u6a5f\u80fd\n* \u97f3\u58f0\u8a55\u4fa1\u30c4\u30fc\u30eb\uff08`scripts/evaluation/`\uff09\n * MCD (Mel-Cepstral Distortion) \u8a55\u4fa1\n * PESQ (Perceptual Evaluation of Speech Quality) \u8a55\u4fa1\n * UTMOS\u8a55\u4fa1\n\n## \u95a2\u9023\u8a18\u4e8b\n* [LJSpeech\u3092\u4f7f\u3063\u3066\u82f1\u8a9e\u306epiper\u306e\u4e8b\u524d\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u4f5c\u6210\u3059\u308b](https://ayousanz.hatenadiary.jp/entry/2025/05/26/230341)\n* [jvs\u97f3\u58f0\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u4f7f\u3063\u305fpiper\u65e5\u672c\u8a9e\u30e2\u30c7\u30eb\u306e\u4f5c\u6210](https://ayousanz.hatenadiary.jp/entry/2025/06/05/093217)\n* [piper\u30e2\u30c7\u30eb\u304b\u3089\u3064\u304f\u3088\u307f\u3061\u3083\u3093\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u4f7f\u3063\u3066\u8ffd\u52a0\u5b66\u7fd2\u3092\u884c\u3046](https://ayousanz.hatenadiary.jp/entry/2025/06/07/074232)\n\n``` sh\necho 'Welcome to the world of speech synthesis!' | \\\n ./piper --model en_US-lessac-medium.onnx --output_file welcome.wav\n```\n\n[Listen to voice samples](https://rhasspy.github.io/piper-samples) and check out a [video tutorial by Thorsten M\u00fcller](https://youtu.be/rjq5eZoWWSo)\n\nVoices are trained with [VITS](https://github.com/jaywalnut310/vits/) and exported to the [onnxruntime](https://onnxruntime.ai/).\n\n[](https://www.openhomefoundation.org/)\n\n## \u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u30b5\u30dd\u30fc\u30c8\n\n### \u5bfe\u5fdc\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\n\n| \u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0 | \u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3 | OpenJTalk\u5bfe\u5fdc | \u5099\u8003 |\n|-----------------|---------------|--------------|------|\n| Linux | x86_64 (amd64) | \u2705 | \u30d5\u30eb\u30b5\u30dd\u30fc\u30c8 |\n| Linux | ARM64 | \u2705 | \u30d5\u30eb\u30b5\u30dd\u30fc\u30c8 (CMake\u30d3\u30eb\u30c9\u4f7f\u7528) |\n| macOS | **ARM64 (Apple Silicon)\u306e\u307f** | \u2705 | M1/M2/M3\u4ee5\u964d\u306eMac\u5c02\u7528 |\n| Windows | x64 | \u2705 | \u30d5\u30eb\u30b5\u30dd\u30fc\u30c8 |\n\n### \u26a0\ufe0f \u91cd\u8981: macOS\u30e6\u30fc\u30b6\u30fc\u3078\u306e\u304a\u77e5\u3089\u305b\n\n**2024\u5e74\u3088\u308a\u3001macOS\u3067\u306fApple Silicon (M1/M2/M3\u4ee5\u964d) \u306e\u307f\u3092\u30b5\u30dd\u30fc\u30c8\u3057\u3066\u3044\u307e\u3059\u3002**\n\n#### Intel Mac\u3092\u304a\u4f7f\u3044\u306e\u65b9\u3078\nIntel Mac (x86_64) \u306e\u30b5\u30dd\u30fc\u30c8\u306f\u7d42\u4e86\u3057\u307e\u3057\u305f\u3002\u4ee5\u4e0b\u306e\u4ee3\u66ff\u65b9\u6cd5\u3092\u3054\u5229\u7528\u304f\u3060\u3055\u3044\uff1a\n\n1. **Docker\u3092\u4f7f\u7528\uff08\u63a8\u5968\uff09**\n ```bash\n # Docker\u30a4\u30e1\u30fc\u30b8\u3092\u30d7\u30eb\n docker pull ghcr.io/ayutaz/piper-plus:latest\n \n # \u5b9f\u884c\u4f8b\n docker run --rm -v $(pwd):/data ghcr.io/ayutaz/piper-plus:latest \\\n echo \"Hello from Docker\" | piper --model /data/model.onnx --output_file /data/output.wav\n ```\n\n2. **\u30bd\u30fc\u30b9\u304b\u3089\u30d3\u30eb\u30c9**\n ```bash\n # \u4f9d\u5b58\u95a2\u4fc2\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\n brew install cmake onnxruntime\n \n # \u30d3\u30eb\u30c9\n git clone https://github.com/ayutaz/piper-plus.git\n cd piper-plus\n mkdir build && cd build\n cmake .. -DCMAKE_BUILD_TYPE=Release\n make -j$(sysctl -n hw.ncpu)\n ```\n\n3. **\u4eee\u60f3\u30de\u30b7\u30f3\u3067Linux\u7248\u3092\u4f7f\u7528**\n - UTM\u3001Parallels Desktop\u3001VMware Fusion\u306a\u3069\u3092\u4f7f\u7528\n\n#### Apple Silicon\u30e6\u30fc\u30b6\u30fc\u306e\u65b9\u3078\n\u901a\u5e38\u901a\u308a\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u3066\u3054\u5229\u7528\u3044\u305f\u3060\u3051\u307e\u3059\u3002\u521d\u56de\u5b9f\u884c\u6642\u306e\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u8b66\u544a\u306b\u3064\u3044\u3066\u306f\u3001\u4ee5\u4e0b\u3092\u3054\u53c2\u7167\u304f\u3060\u3055\u3044\u3002\n\n##### macOS\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u8b66\u544a\u306e\u5bfe\u51e6\n\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u305f\u30d0\u30a4\u30ca\u30ea\u3092\u521d\u3081\u3066\u5b9f\u884c\u3059\u308b\u969b\u3001macOS\u306e\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u6a5f\u80fd\u306b\u3088\u308a\u8b66\u544a\u304c\u8868\u793a\u3055\u308c\u308b\u5834\u5408\u304c\u3042\u308a\u307e\u3059\u3002\u4ee5\u4e0b\u306e\u30b3\u30de\u30f3\u30c9\u3067\u691c\u75ab\u5c5e\u6027\u3092\u524a\u9664\u3057\u3066\u304f\u3060\u3055\u3044\uff1a\n\n```bash\n# \u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u305f\u30d5\u30a1\u30a4\u30eb\u3092\u5c55\u958b\u5f8c\nxattr -cr piper/\n\n# \u307e\u305f\u306f\u7279\u5b9a\u306e\u30d0\u30a4\u30ca\u30ea\u306e\u307f\nxattr -cr piper/bin/piper\nxattr -cr piper/bin/open_jtalk # \u65e5\u672c\u8a9eTTS\u3092\u4f7f\u7528\u3059\u308b\u5834\u5408\n```\n\n\u3053\u308c\u306b\u3088\u308a\u3001Gatekeeper\u306e\u8b66\u544a\u306a\u3057\u306b\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308a\u307e\u3059\u3002\n\n## Voices\n\nOur goal is to support Home Assistant and the [Year of Voice](https://www.home-assistant.io/blog/2022/12/20/year-of-voice/).\n\n[Download voices](VOICES.md) for the supported languages:\n\n* \u0627\u0644\u0639\u0631\u0628\u064a\u0629, Jordan (Arabic, ar_JO)\n* Catal\u00e0, Spain (Catalan, ca_ES)\n* \u010ce\u0161tina, Czech Republic (Czech, cs_CZ)\n* Cymraeg, Great Britain (Welsh, cy_GB)\n* Dansk, Denmark (Danish, da_DK)\n* Deutsch, Germany (German, de_DE)\n* \u0395\u03bb\u03bb\u03b7\u03bd\u03b9\u03ba\u03ac, Greece (Greek, el_GR)\n* English, Great Britain (English, en_GB)\n* English, United States (English, en_US)\n* Espa\u00f1ol, Argentina (Spanish, es_AR)\n* Espa\u00f1ol, Spain (Spanish, es_ES)\n* Espa\u00f1ol, Mexico (Spanish, es_MX)\n* \u0641\u0627\u0631\u0633\u06cc, Iran (Farsi, fa_IR)\n* Suomi, Finland (Finnish, fi_FI)\n* Fran\u00e7ais, France (French, fr_FR)\n* Magyar, Hungary (Hungarian, hu_HU)\n* \u00edslenska, Iceland (Icelandic, is_IS)\n* Italiano, Italy (Italian, it_IT)\n* \u10e5\u10d0\u10e0\u10d7\u10e3\u10da\u10d8 \u10d4\u10dc\u10d0, Georgia (Georgian, ka_GE)\n* \u049b\u0430\u0437\u0430\u049b\u0448\u0430, Kazakhstan (Kazakh, kk_KZ)\n* L\u00ebtzebuergesch, Luxembourg (Luxembourgish, lb_LU)\n* Latvie\u0161u, Latvia (Latvian, lv_LV)\n* \u0d2e\u0d32\u0d2f\u0d3e\u0d33\u0d02, India (Malayalam, ml_IN)\n* \u0939\u093f\u0902\u0926\u0940, India (Hindi, hi_IN)\n* \u0928\u0947\u092a\u093e\u0932\u0940, Nepal (Nepali, ne_NP)\n* Nederlands, Belgium (Dutch, nl_BE)\n* Nederlands, Netherlands (Dutch, nl_NL)\n* Norsk, Norway (Norwegian, no_NO)\n* Polski, Poland (Polish, pl_PL)\n* Portugu\u00eas, Brazil (Portuguese, pt_BR)\n* Portugu\u00eas, Portugal (Portuguese, pt_PT)\n* Rom\u00e2n\u0103, Romania (Romanian, ro_RO)\n* \u0420\u0443\u0441\u0441\u043a\u0438\u0439, Russia (Russian, ru_RU)\n* Sloven\u010dina, Slovakia (Slovak, sk_SK)\n* Sloven\u0161\u010dina, Slovenia (Slovenian, sl_SI)\n* srpski, Serbia (Serbian, sr_RS)\n* Svenska, Sweden (Swedish, sv_SE)\n* Kiswahili, Democratic Republic of the Congo (Swahili, sw_CD)\n* T\u00fcrk\u00e7e, Turkey (Turkish, tr_TR)\n* \u0443\u043a\u0440\u0430\u0457\u0301\u043d\u0441\u044c\u043a\u0430 \u043c\u043e\u0301\u0432\u0430, Ukraine (Ukrainian, uk_UA)\n* Ti\u1ebfng Vi\u1ec7t, Vietnam (Vietnamese, vi_VN)\n* \u7b80\u4f53\u4e2d\u6587, China (Chinese, zh_CN)\n\nYou will need two files per voice:\n\n1. A `.onnx` model file, such as [`en_US-lessac-medium.onnx`](https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/en/en_US/lessac/medium/en_US-lessac-medium.onnx)\n2. A `.onnx.json` config file, such as [`en_US-lessac-medium.onnx.json`](https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/en/en_US/lessac/medium/en_US-lessac-medium.onnx.json)\n\nThe `MODEL_CARD` file for each voice contains important licensing information. Piper is intended for text to speech research, and does not impose any additional restrictions on voice models. Some voices may have restrictive licenses, however, so please review them carefully!\n\n\n## Installation\n\nYou can [run Piper with Python](#running-in-python) or download a binary release:\n\n* [amd64](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_amd64.tar.gz) (64-bit desktop Linux)\n* [arm64](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_arm64.tar.gz) (64-bit Raspberry Pi 4)\n* [armv7](https://github.com/rhasspy/piper/releases/download/v1.2.0/piper_armv7.tar.gz) (32-bit Raspberry Pi 3/4)\n\n### Building from Source\n\nIf you want to build from source, see the [Makefile](Makefile) and [C++ source](src/cpp).\n\n#### Prerequisites\n\n* C++ compiler with C++17 support\n* CMake 3.13 or later\n* Git\n\n#### Build Steps\n\n1. Clone the repository:\n ```bash\n git clone https://github.com/rhasspy/piper.git\n cd piper\n ```\n\n2. Create build directory:\n ```bash\n mkdir build\n cd build\n ```\n\n3. Configure and build:\n ```bash\n cmake ..\n cmake --build . --config Release\n ```\n\n#### Platform-specific Notes\n\n**Linux**: You must download and extract [piper-phonemize](https://github.com/rhasspy/piper-phonemize) to `lib/Linux-$(uname -m)/piper_phonemize` before building.\nFor example, `lib/Linux-x86_64/piper_phonemize/lib/libpiper_phonemize.so` should exist for AMD/Intel machines.\n\n**Windows**: See the [Windows Setup Guide](docs/windows-setup.md) for detailed instructions.\n\n**macOS**: The build process will automatically download required dependencies.\n\n\n## Usage\n\n1. [Download a voice](#voices) and extract the `.onnx` and `.onnx.json` files\n2. Run the `piper` binary with text on standard input, `--model /path/to/your-voice.onnx`, and `--output_file output.wav`\n\nFor example:\n\n``` sh\necho 'Welcome to the world of speech synthesis!' | \\\n ./piper --model en_US-lessac-medium.onnx --output_file welcome.wav\n```\n\nFor multi-speaker models, use `--speaker <number>` to change speakers (default: 0).\n\n### Additional Options\n\n* `--use-cuda` - Enable GPU acceleration with CUDA\n* `--quiet` / `-q` - Disable logging output\n* `--phoneme-silence <phoneme> <seconds>` - Set silence duration for specific phonemes\n* `--length-scale <value>` - Adjust speech speed (default: 1.0, smaller = faster)\n* `--noise-scale <value>` - Control audio variation (default: 0.667)\n* `--noise-w <value>` - Control phoneme duration variation (default: 0.8)\n* `--sentence-silence <seconds>` - Silence between sentences (default: 0.2)\n\nSee `piper --help` for more options.\n\n### Streaming Audio\n\nPiper can stream raw audio to stdout as its produced:\n\n``` sh\necho 'This sentence is spoken first. This sentence is synthesized while the first sentence is spoken.' | \\\n ./piper --model en_US-lessac-medium.onnx --output-raw | \\\n aplay -r 22050 -f S16_LE -t raw -\n```\n\nThis is **raw** audio and not a WAV file, so make sure your audio player is set to play 16-bit mono PCM samples at the correct sample rate for the voice.\n\n### JSON Input\n\nThe `piper` executable can accept JSON input when using the `--json-input` flag. Each line of input must be a JSON object with `text` field. For example:\n\n``` json\n{ \"text\": \"First sentence to speak.\" }\n{ \"text\": \"Second sentence to speak.\" }\n```\n\nOptional fields include:\n\n* `speaker` - string\n * Name of the speaker to use from `speaker_id_map` in config (multi-speaker voices only)\n* `speaker_id` - number\n * Id of speaker to use from 0 to number of speakers - 1 (multi-speaker voices only, overrides \"speaker\")\n* `output_file` - string\n * Path to output WAV file\n \nThe following example writes two sentences with different speakers to different files:\n\n``` json\n{ \"text\": \"First speaker.\", \"speaker_id\": 0, \"output_file\": \"/tmp/speaker_0.wav\" }\n{ \"text\": \"Second speaker.\", \"speaker_id\": 1, \"output_file\": \"/tmp/speaker_1.wav\" }\n```\n\n\n## People using Piper\n\nPiper has been used in the following projects/papers:\n\n* [Home Assistant](https://github.com/home-assistant/addons/blob/master/piper/README.md)\n* [Rhasspy 3](https://github.com/rhasspy/rhasspy3/)\n* [NVDA - NonVisual Desktop Access](https://www.nvaccess.org/post/in-process-8th-may-2023/#voices)\n* [Image Captioning for the Visually Impaired and Blind: A Recipe for Low-Resource Languages](https://www.techrxiv.org/articles/preprint/Image_Captioning_for_the_Visually_Impaired_and_Blind_A_Recipe_for_Low-Resource_Languages/22133894)\n* [Open Voice Operating System](https://github.com/OpenVoiceOS/ovos-tts-plugin-piper)\n* [JetsonGPT](https://github.com/shahizat/jetsonGPT)\n* [LocalAI](https://github.com/go-skynet/LocalAI)\n* [Lernstick EDU / EXAM: reading clipboard content aloud with language detection](https://lernstick.ch/)\n* [Natural Speech - A plugin for Runelite, an OSRS Client](https://github.com/phyce/rl-natural-speech)\n* [mintPiper](https://github.com/evuraan/mintPiper)\n* [Vim-Piper](https://github.com/wolandark/vim-piper)\n\n## Training\n\nSee the [training guide](TRAINING.md) and the [source code](src/python).\n\nPretrained checkpoints are available on [Hugging Face](https://huggingface.co/datasets/rhasspy/piper-checkpoints/tree/main)\n\n\n## Running in Python\n\nSee [src/python_run](src/python_run)\n\nInstall with `pip`:\n\n``` sh\n# \u57fa\u672c\u6a5f\u80fd\u306e\u307f\npip install piper-tts-plus\n\n# GPU \u7248 (CUDA \u74b0\u5883\u304c\u3042\u308b\u5834\u5408)\npip install \"piper-tts-plus[gpu]\"\n\n# HTTP \u30b5\u30fc\u30d0\u30fc\u6a5f\u80fd\u3092\u542b\u3080\u5834\u5408\npip install \"piper-tts-plus[http]\"\n\n# GPU + HTTP\npip install \"piper-tts-plus[gpu,http]\"\n```\n\nThis will automatically download [voice files](https://huggingface.co/rhasspy/piper-voices/tree/v1.0.0) the first time they're used. Use `--data-dir` and `--download-dir` to adjust where voices are found/downloaded.\n\nIf you'd like to use a GPU, install the `onnxruntime-gpu` package:\n\n\n``` sh\n.venv/bin/pip3 install onnxruntime-gpu\n```\n\nand then run `piper` with the `--cuda` argument. You will need to have a functioning CUDA environment, such as what's available in [NVIDIA's PyTorch containers](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch).\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A fast, local neural text to speech system that sounds great and is optimized for the Raspberry Pi 4.",
"version": "1.3.0",
"project_urls": {
"Bug Reports": "https://github.com/ayutaz/piper-plus/issues",
"Documentation": "https://github.com/ayutaz/piper-plus/blob/dev/README.md",
"Homepage": "https://github.com/ayutaz/piper-plus",
"Source": "https://github.com/ayutaz/piper-plus"
},
"split_keywords": [
"piper-plus",
"tts",
"text-to-speech",
"japanese",
"openjtalk",
"neural-tts",
"vits",
"multi-gpu",
"pytorch-lightning",
"onnx",
"raspberry-pi"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "95f31779849f663b959fdc8b84bf9d3231d957ddc2b922f98fed44cd2239fe9c",
"md5": "65dde1c8dcc64cd9b44d3675b323f892",
"sha256": "f89dadd14936e83d512c3afef4c67c1cce0de5514fed82cffe75c1296108479d"
},
"downloads": -1,
"filename": "piper_tts_plus-1.3.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "65dde1c8dcc64cd9b44d3675b323f892",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.11",
"size": 43767,
"upload_time": "2025-07-20T21:16:27",
"upload_time_iso_8601": "2025-07-20T21:16:27.212731Z",
"url": "https://files.pythonhosted.org/packages/95/f3/1779849f663b959fdc8b84bf9d3231d957ddc2b922f98fed44cd2239fe9c/piper_tts_plus-1.3.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "464ff4f2b6dd28abcf699f2145d7a958afc361cd4f9ce678bda3a3e8b659f6e2",
"md5": "2e7e941c60b7f30933b963fbb117e3e9",
"sha256": "0af01b87fb09d6464bc3741389905bffa61ee955b57304e5462381aa48728251"
},
"downloads": -1,
"filename": "piper_tts_plus-1.3.0.tar.gz",
"has_sig": false,
"md5_digest": "2e7e941c60b7f30933b963fbb117e3e9",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.11",
"size": 49322,
"upload_time": "2025-07-20T21:16:29",
"upload_time_iso_8601": "2025-07-20T21:16:29.021741Z",
"url": "https://files.pythonhosted.org/packages/46/4f/f4f2b6dd28abcf699f2145d7a958afc361cd4f9ce678bda3a3e8b659f6e2/piper_tts_plus-1.3.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-07-20 21:16:29",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "ayutaz",
"github_project": "piper-plus",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "piper-tts-plus"
}
JSON
