Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
docs: clarify mixed-device source version requirement
zhifu gaob2d63f021aa9
diff --git a/docs/python_api.md b/docs/python_api.mdindex c4b3c72c0..d05a6ad7c 100644--- a/docs/python_api.md+++ b/docs/python_api.md@@ -53,6 +53,8 @@ The wrapper accepts `**kwargs`, not a universally validated parameter schema. A### Place submodels on separate devices+> **Version requirement:** This placement contract is not included in the PyPI 1.4.15 package. Use a source checkout containing [PR #3706](https://github.com/modelscope/FunASR/pull/3706), merged at [4d8e087](https://github.com/modelscope/FunASR/commit/4d8e08748c9d7cc10f3e3a01f5f4b927a8948699), and follow the [source installation instructions](installation/installation.md). Record the checkout commit and imported `funasr.__file__`; a development checkout can still report version `1.4.15`.+For example, keep FSMN VAD and punctuation on CPU while ASR uses Apple Silicon MPS:```pythondiff --git a/docs/python_api_zh.md b/docs/python_api_zh.mdindex 5128c16a4..36e658d69 100644--- a/docs/python_api_zh.md+++ b/docs/python_api_zh.md@@ -53,6 +53,8 @@### 为子模型分别指定设备+> **版本要求:** PyPI 的 1.4.15 发行包尚不包含此设备配置契约。请使用包含 [PR #3706](https://github.com/modelscope/FunASR/pull/3706) 的源码版本(合并提交为 [4d8e087](https://github.com/modelscope/FunASR/commit/4d8e08748c9d7cc10f3e3a01f5f4b927a8948699)),并按照[源码安装说明](installation/installation_zh.md)操作。记录工作区 commit 和实际导入的 `funasr.__file__`;开发版工作区仍可能显示版本号 `1.4.15`。+例如,在 Apple Silicon 上让 ASR 使用 MPS,而 FSMN VAD 和标点使用 CPU:```python