canto-tts 0.1.0__tar.gz

canto_tts-0.1.0/.gitignore

@@ -0,0 +1,113 @@
+# =============================================================================
+# canto-tts — .gitignore
+# =============================================================================
+
+# ---------------------------------------------------------------------------
+# Python bytecode & caches
+# ---------------------------------------------------------------------------
+__pycache__/
+*.py[cod]
+*$py.class
+*.pyo
+
+# ---------------------------------------------------------------------------
+# Virtual environments
+# ---------------------------------------------------------------------------
+.venv/
+venv/
+env/
+ENV/
+.env/
+
+# ---------------------------------------------------------------------------
+# Build artefacts & distribution
+# ---------------------------------------------------------------------------
+build/
+dist/
+*.egg-info/
+*.egg
+MANIFEST
+.eggs/
+
+# ---------------------------------------------------------------------------
+# Installer / package manager state
+# ---------------------------------------------------------------------------
+pip-log.txt
+pip-delete-this-directory.txt
+*.whl
+.uv/
+
+# ---------------------------------------------------------------------------
+# Testing & coverage
+# ---------------------------------------------------------------------------
+.pytest_cache/
+.coverage
+.coverage.*
+htmlcov/
+coverage.xml
+*.cover
+
+# ---------------------------------------------------------------------------
+# Type checking
+# ---------------------------------------------------------------------------
+.mypy_cache/
+.dmypy.json
+.pytype/
+.pyre/
+
+# ---------------------------------------------------------------------------
+# Linting
+# ---------------------------------------------------------------------------
+.ruff_cache/
+
+# ---------------------------------------------------------------------------
+# IDEs & editors
+# ---------------------------------------------------------------------------
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+.DS_Store
+Thumbs.db
+
+# ---------------------------------------------------------------------------
+# Local model / weight cache directories
+# (never commit large binary weight files to the source repo)
+# ---------------------------------------------------------------------------
+onnx_weights/
+torch_weights/
+model_cache/
+weights/
+checkpoints/
+
+# ---------------------------------------------------------------------------
+# Audio test outputs
+# (generated by tests or manual runs — not source artefacts)
+# ---------------------------------------------------------------------------
+*.wav
+*.mp3
+*.flac
+*.ogg
+*.opus
+
+# ---------------------------------------------------------------------------
+# Jupyter / notebook
+# ---------------------------------------------------------------------------
+.ipynb_checkpoints/
+*.ipynb
+
+# ---------------------------------------------------------------------------
+# Docs build output
+# ---------------------------------------------------------------------------
+docs/_build/
+docs/site/
+site/
+
+# ---------------------------------------------------------------------------
+# Misc
+# ---------------------------------------------------------------------------
+*.log
+*.tmp
+*.bak
+.hypothesis/

canto_tts-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

canto_tts-0.1.0/PKG-INFO

@@ -0,0 +1,120 @@
+Metadata-Version: 2.4
+Name: canto-tts
+Version: 0.1.0
+Summary: Open-source Cantonese (Hong Kong) text-to-speech — CPU-first ONNX inference SDK
+Project-URL: Homepage, https://github.com/typangaa/canto-tts
+Project-URL: Repository, https://github.com/typangaa/canto-tts
+Author: typangaa
+License: Apache-2.0
+License-File: LICENSE
+Requires-Python: >=3.10
+Requires-Dist: canto-hk-g2p
+Requires-Dist: huggingface-hub>=0.20
+Requires-Dist: numpy
+Requires-Dist: onnxruntime>=1.16
+Requires-Dist: sentencepiece
+Requires-Dist: soundfile
+Provides-Extra: all
+Requires-Dist: fastapi>=0.110; extra == 'all'
+Requires-Dist: faster-whisper>=1.0; extra == 'all'
+Requires-Dist: funasr; extra == 'all'
+Requires-Dist: jinja2>=3.1; extra == 'all'
+Requires-Dist: modelscope; extra == 'all'
+Requires-Dist: opencc-python-reimplemented; extra == 'all'
+Requires-Dist: python-multipart; extra == 'all'
+Requires-Dist: torch>=2.0; extra == 'all'
+Requires-Dist: torchaudio>=2.0; extra == 'all'
+Requires-Dist: transformers==4.57.1; extra == 'all'
+Requires-Dist: uvicorn>=0.27; extra == 'all'
+Provides-Extra: demo
+Requires-Dist: fastapi>=0.110; extra == 'demo'
+Requires-Dist: jinja2>=3.1; extra == 'demo'
+Requires-Dist: python-multipart; extra == 'demo'
+Requires-Dist: uvicorn>=0.27; extra == 'demo'
+Provides-Extra: quality
+Requires-Dist: faster-whisper>=1.0; extra == 'quality'
+Provides-Extra: quality-sensevoice
+Requires-Dist: funasr; extra == 'quality-sensevoice'
+Requires-Dist: modelscope; extra == 'quality-sensevoice'
+Requires-Dist: opencc-python-reimplemented; extra == 'quality-sensevoice'
+Requires-Dist: torch>=2.0; extra == 'quality-sensevoice'
+Requires-Dist: torchaudio>=2.0; extra == 'quality-sensevoice'
+Provides-Extra: torch
+Requires-Dist: torch>=2.0; extra == 'torch'
+Requires-Dist: torchaudio>=2.0; extra == 'torch'
+Requires-Dist: transformers==4.57.1; extra == 'torch'
+Description-Content-Type: text/markdown
+
+# canto-tts
+
+**Cantonese Hong Kong Text-to-Speech · CPU-first · Apache-2.0**
+
+[![PyPI](https://img.shields.io/pypi/v/canto-tts)](https://pypi.org/project/canto-tts/)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue)](https://github.com/typangaa/canto-tts/blob/main/LICENSE)
+[![Python](https://img.shields.io/pypi/pyversions/canto-tts)](https://pypi.org/project/canto-tts/)
+
+An open-source Cantonese (Hong Kong) text-to-speech SDK.
+Runs on CPU via ONNX Runtime — no GPU required.
+Fine-tuned from [MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS)
+(0.1 B params, GPT-2 backbone, Apache-2.0, by OpenMOSS).
+
+> **Weights published** — CER 11.82% / tone accuracy 84.22% / code-switch CER 13.87%
+> on a 100-sentence gate set (N=5 repeat eval).
+> Single default voice only — no voice cloning, no voice selection (yet).
+> **PyPI package not published yet** — install from source for now;
+> see the [GitHub repo](https://github.com/typangaa/canto-tts) for current status.
+
+---
+
+## Install
+
+```bash
+git clone https://github.com/typangaa/canto-tts.git && cd canto-tts && pip install -e .
+```
+
+## Quickstart
+
+```python
+from canto_tts import CantoTTS
+
+tts = CantoTTS()  # auto-downloads typangaa/canto-tts-nano from HuggingFace on first use
+tts.synthesize("多謝晒，今日天氣幾好。", "hello.wav")
+print("Saved to hello.wav")
+
+# English code-switching works too
+tts.synthesize("我哋一齊去 IFC food court 食飯。", "codeswitching.wav")
+```
+
+Weights auto-download from HuggingFace and cache via the Hub (`~/.cache/huggingface/hub/`).
+
+## Quality Modes (opt-in)
+
+Two opt-in `quality=` modes trade extra compute for a more reliable draw, without touching the model:
+
+```python
+tts.synthesize(text, "out.wav", quality="duration_filter", max_attempts=3)   # no extra deps
+tts.synthesize(text, "out.wav", quality="best_of_n", best_of_n=4)            # needs canto-tts[quality]
+```
+
+`quality="best_of_n"` reranks N draws by ASR character-error-rate; pass `asr_backend="sensevoice"` (needs `canto-tts[quality-sensevoice]`) for ~7x faster reranking at a small accuracy cost. Full writeup: [GitHub README](https://github.com/typangaa/canto-tts#quality-modes-opt-in-inference-time-reranking).
+
+## Limitations
+
+| | |
+|--|--|
+| **Quality** | CER 11.82% / tone accuracy 84.22% / code-switch CER 13.87% (N=5 repeat, 100-sentence gate set). |
+| **Voice** | Single default voice. No voice cloning, no voice selection (yet). |
+| **Language** | Cantonese (Hong Kong) + English code-switching only. No Mandarin. |
+| **Training data** | Privately sourced — not released (copyright), regardless of weight status. |
+| **Weights** | ✅ Published — [huggingface.co/typangaa/canto-tts-nano](https://huggingface.co/typangaa/canto-tts-nano). |
+
+## Model & License
+
+- **Base model**: [MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS) by OpenMOSS
+- **Weights**: published at [huggingface.co/typangaa/canto-tts-nano](https://huggingface.co/typangaa/canto-tts-nano)
+- **License**: Apache-2.0 (inherited from MOSS-TTS-Nano)
+
+## Full Documentation
+
+Full README, CLI docs, Docker self-hosting, and deployment guide:
+**https://github.com/typangaa/canto-tts**

canto_tts-0.1.0/README.md

@@ -0,0 +1,206 @@
+# canto-tts 🎙️
+
+> **廣東話（香港）文字轉語音 · Cantonese Hong Kong Text-to-Speech**
+> CPU-first, ONNX Runtime, Apache-2.0 — no GPU required.
+
+[![PyPI](https://img.shields.io/pypi/v/canto-tts)](https://pypi.org/project/canto-tts/)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue)](LICENSE)
+[![Python](https://img.shields.io/pypi/pyversions/canto-tts)](https://pypi.org/project/canto-tts/)
+[![HuggingFace](https://img.shields.io/badge/🤗%20Model-typangaa%2Fcanto--tts--nano-yellow)](https://huggingface.co/typangaa/canto-tts-nano)
+
+---
+
+## ✅ 狀態 / Status
+
+> **Weights published on HuggingFace** — [`typangaa/canto-tts-nano`](https://huggingface.co/typangaa/canto-tts-nano).
+> `canto-tts-nano-v1` 已經通過晒全部 quality gate(CER 11.82% / tone accuracy 84.22% / code-switch CER
+> 13.87%,N=5 repeat eval),`CantoTTS()` 而家可以零參數自動下載呢個 checkpoint。**PyPI package 未發佈**——
+> 而家請用 [Install](#install) 嘅 source install。
+>
+> Quality gate 數字全表見 [Limitations](#limitations--已知限制) / [Model & License](#model--license)。
+
+See the [Limitations](#limitations--已知限制) section for full details.
+
+---
+
+## 廣東話簡介
+
+`canto-tts` 係一個開源嘅廣東話（香港）文字轉語音 Python SDK，底層用 ONNX Runtime，唔需要 GPU。模型係由 [MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS) (OpenMOSS, Apache-2.0) fine-tune 而成，支援廣東話拼音輸入（jyutping phoneme tokens），以及英文 code-switching（英文字保留原文，唔需要另外做音素轉換）。
+
+---
+
+## English Description
+
+`canto-tts` is an open-source Cantonese (Hong Kong) text-to-speech Python SDK.
+It runs on CPU via ONNX Runtime (no GPU needed) and is fine-tuned from
+[MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS) (0.1 B parameters, GPT-2 backbone, Apache-2.0).
+
+Key facts:
+- **Input**: jyutping-phoneme tokens (the library converts raw Hanzi for you via `canto_hk_g2p`).
+- **English code-switching**: English words inside a Cantonese sentence are kept as orthography and pronounced naturally.
+- **CPU-first**: default backend is ONNX Runtime; PyTorch is optional.
+- **Single default voice**: unconditional generation — no voice cloning, no voice selection.
+- **Weights**: auto-downloads from [typangaa/canto-tts-nano](https://huggingface.co/typangaa/canto-tts-nano) on first use (cached under `~/.cache/huggingface`) — see [Status](#-狀態--status) above.
+- **Training data**: privately sourced (not released for copyright reasons) — only the model weights will be public (once quality gates are met).
+
+---
+
+## Install
+
+⚠️ **Not on PyPI yet.** Install from source for now:
+
+```bash
+git clone https://github.com/typangaa/canto-tts.git
+cd canto-tts
+pip install -e .
+```
+
+Once published, this will be `pip install canto-tts`.
+
+Optional extras:
+
+| Extra | What it adds |
+|-------|-------------|
+| `canto-tts[demo]` | FastAPI demo server (`canto-tts-demo` command) |
+| `canto-tts[torch]` | PyTorch backend |
+| `canto-tts[quality]` | `quality="best_of_n"` with the default ASR reranker (torch-free) — see [Quality Modes](#quality-modes-opt-in-inference-time-reranking) |
+| `canto-tts[quality-sensevoice]` | `quality="best_of_n"` with the faster `asr_backend="sensevoice"` reranker (adds torch) |
+| `canto-tts[dev]`  | Dev / test tools |
+
+---
+
+## Quickstart (Python SDK)
+
+```python
+from canto_tts import CantoTTS
+
+tts = CantoTTS()  # auto-downloads typangaa/canto-tts-nano from HuggingFace on first use
+tts.synthesize("多謝晒，今日天氣幾好。", "hello.wav")
+print("Saved to hello.wav")
+
+# Code-switching: English words inside Cantonese are read naturally
+tts.synthesize("我哋一齊去 IFC food court 食飯。", "codeswitching.wav")
+```
+
+To point at your own locally-exported ONNX bundle instead (e.g. a checkpoint you fine-tuned
+yourself via [`scripts/export_onnx.py`](scripts/export_onnx.py)), pass `checkpoint=`:
+
+```python
+tts = CantoTTS(checkpoint="/path/to/your/exported/onnx_weights")
+```
+
+See [`examples/quickstart.py`](examples/quickstart.py) for the full annotated version.
+
+---
+
+## CLI
+
+```bash
+canto-tts synthesize "多謝晒，今日天氣幾好。" -o hello.wav   # auto-downloads weights on first use
+
+# Point at your own local export instead
+canto-tts synthesize "..." -o out.wav --backend onnx --checkpoint /path/to/onnx_weights
+```
+
+Run `canto-tts --help` for all commands.
+
+---
+
+## Quality Modes (opt-in inference-time reranking)
+
+`synthesize()` 預設淨係跑一次 draw(最快)。兩種 opt-in `quality=` mode 都唔改 model 本身,淨係揀邊個 draw 好過留低:
+
+- `quality="duration_filter"`:最多跑 `max_attempts`(default 3)次,揀 duration 最貼近 phoneme-length 預期嘅一個。捕捉最常見嘅兩種 catastrophic AR-codec 失敗(過早截斷 / 無限循環)。**唔使裝額外依賴**,一停到 in-range 嘅 draw 就即刻停,唔一定跑晒個 budget。
+- `quality="best_of_n"`:跑 `best_of_n`(default 4)次,每次用本地 ASR model 轉錄,揀同輸入文字 character-error-rate(CER)最低嘅一個。捕捉 duration 篩唔到嘅問題(錯調、發音錯、code-switch 段落含糊)。需要 `canto-tts[quality]`(`asr_backend="whisper"`,default)或 `canto-tts[quality-sensevoice]`(`asr_backend="sensevoice"`)。
+
+```python
+tts.synthesize(text, "out.wav", quality="duration_filter", max_attempts=3)
+tts.synthesize(text, "out.wav", quality="best_of_n", best_of_n=4, asr_backend="whisper")
+```
+
+```bash
+canto-tts synthesize "..." -o out.wav --quality best_of_n --best-of-n 4 --asr-backend whisper
+```
+
+`asr_backend` options for `quality="best_of_n"` (measured on this project's own generation output — mean CER = the ASR's own transcription error vs. known text, i.e. how sharp a reranking signal it gives, not a general-purpose ASR quality claim):
+
+| `asr_backend` | Extra | Mean CER | Speed | Dependency footprint |
+|---|---|---|---|---|
+| `"whisper"` (default) | `canto-tts[quality]` | 0.036 (most accurate) | ~1.3s/candidate | torch-free (faster-whisper / CTranslate2, a Cantonese fine-tune of whisper-small) |
+| `"sensevoice"` | `canto-tts[quality-sensevoice]` | 0.053 | ~0.18s/candidate (~7x faster) | pulls in torch + torchaudio; non-OSI ModelScope model license (commercial use permitted) |
+
+See [`canto_tts/quality.py`](src/canto_tts/quality.py)'s module docstring for the full tradeoff writeup.
+
+---
+
+## Web Demo (self-hosted)
+
+```bash
+pip install -e ".[demo]"
+canto-tts-demo   # auto-downloads weights on first use; set CANTO_TTS_CHECKPOINT to override
+# → open http://localhost:8000
+```
+
+The demo provides a browser-based UI with a text input and an audio player.
+No API key required — designed for local / self-hosted use.
+
+---
+
+## Docker
+
+```bash
+# weights auto-download on first use; only edit docker/docker-compose.yml if you
+# want to bind-mount your own local export and set CANTO_TTS_CHECKPOINT instead
+docker compose -f docker/docker-compose.yml up
+# → open http://localhost:8000
+```
+
+See [`docs/DEPLOY.md`](docs/DEPLOY.md) for full self-hosting instructions.
+
+---
+
+## Limitations / 已知限制
+
+| Item | Detail |
+|------|--------|
+| **Quality** | CER 11.82% / tone accuracy 84.22% / code-switch CER 13.87% on a 100-sentence gate set (N=5 repeat eval, Qwen3-ASR judge). Pure-English CER 9.7%. Voice-clone (zero-shot) CER 10.2%. |
+| **Voice** | Single baked default voice in the published ONNX bundle. No runtime voice selection or cloning via the ONNX backend; the underlying checkpoint supports zero-shot voice cloning (see the torch backend / `--ref-audio`), a multi-voice picker for the ONNX path is planned. |
+| **Language** | Cantonese (Hong Kong) + English code-switching only. **No Mandarin support** (intentionally not preserved). |
+| **Training data** | Privately sourced — not released (copyright), regardless of weight-release status. |
+| **Weights** | ✅ Published — [huggingface.co/typangaa/canto-tts-nano](https://huggingface.co/typangaa/canto-tts-nano). |
+| **Input** | Converts Hanzi → jyutping internally via `canto_hk_g2p`. Homophones are disambiguated by the G2P model; errors are possible. |
+| **Audio** | 48 000 Hz stereo WAV output (native codec rate — see `MOSS-Audio-Tokenizer-Nano`). |
+
+---
+
+## Model & License
+
+| | |
+|--|--|
+| **Base model** | [MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS) by OpenMOSS — 0.1 B params, GPT-2 backbone |
+| **Fine-tune** | Hong Kong Cantonese, privately sourced training data |
+| **Weights** | ✅ Published at [huggingface.co/typangaa/canto-tts-nano](https://huggingface.co/typangaa/canto-tts-nano) |
+| **License** | [Apache-2.0](LICENSE) (inherited from MOSS-TTS-Nano) |
+| **GitHub** | [github.com/typangaa/canto-tts](https://github.com/typangaa/canto-tts) |
+
+---
+
+## Contributing
+
+Issues and PRs welcome. Please open an issue before starting significant work.
+See `CONTRIBUTING.md` (coming soon) for coding style and test conventions.
+
+---
+
+## Citation
+
+If you use this work in research, please cite the base model:
+
+```bibtex
+@misc{moss-tts-nano,
+  author    = {OpenMOSS},
+  title     = {MOSS-TTS-Nano},
+  year      = {2024},
+  url       = {https://github.com/OpenMOSS/MOSS-TTS}
+}
+```