|
1 | | -## rapid_paraformer |
2 | | - |
3 | | -<p align="left"> |
4 | | - <a href=""><img src="https://img.shields.io/badge/OS-Linux%2C%20Win%2C%20Mac-pink.svg"></a> |
5 | | - <a href=""><img src="https://img.shields.io/badge/Python->=3.6,<3.12-aff.svg"></a> |
6 | | - <a href="https://pepy.tech/project/rapid_paraformer"><img src="https://static.pepy.tech/personalized-badge/rapid_paraformer?period=total&units=abbreviation&left_color=grey&right_color=blue&left_text=Downloads"></a> |
7 | | - <a href="https://pypi.org/project/rapid_paraformer/"><img alt="PyPI" src="https://img.shields.io/pypi/v/rapid_paraformer"></a> |
8 | | - <a href="https://semver.org/"><img alt="SemVer2.0" src="https://img.shields.io/badge/SemVer-2.0-brightgreen"></a> |
9 | | - <a href="https://github.com/psf/black"><img src="https://img.shields.io/badge/code%20style-black-000000.svg"></a> |
10 | | -</p> |
11 | | - |
12 | | -- 模型出自阿里达摩院[Paraformer语音识别-中文-通用-16k-离线-large-pytorch](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) |
13 | | -- 🎉该项目核心代码已经并入[FunASR](https://github.com/alibaba-damo-academy/FunASR) |
14 | | -- 本仓库仅对模型做了转换，只采用ONNXRuntime推理引擎 |
15 | | - |
16 | | -#### TODO |
| 1 | +<div align="center"> |
| 2 | +<div align="center"> |
| 3 | + <h1><b>Rapid Paraformer</b></h1> |
| 4 | +</div> |
| 5 | + |
| 6 | +<a href=""><img src="https://img.shields.io/badge/OS-Linux%2C%20Win%2C%20Mac-pink.svg"></a> |
| 7 | +<a href=""><img src="https://img.shields.io/badge/Python->=3.6,<3.13-aff.svg"></a> |
| 8 | +<a href="https://pepy.tech/project/rapid_paraformer"><img src="https://static.pepy.tech/personalized-badge/rapid_paraformer?period=total&units=abbreviation&left_color=grey&right_color=blue&left_text=Downloads"></a> |
| 9 | +<a href="https://pypi.org/project/rapid_paraformer/"><img alt="PyPI" src="https://img.shields.io/pypi/v/rapid_paraformer"></a> |
| 10 | +<a href="https://semver.org/"><img alt="SemVer2.0" src="https://img.shields.io/badge/SemVer-2.0-brightgreen"></a> |
| 11 | +<a href="https://github.com/psf/black"><img src="https://img.shields.io/badge/code%20style-black-000000.svg"></a> |
| 12 | +</div> |
| 13 | + |
| 14 | +## 简介 |
| 15 | +rapid_paraformer是一个基于阿里达摩院[Paraformer语音识别-中文-通用-16k-离线-large-pytorch](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary)的语音识别工具。 |
| 16 | + |
| 17 | +🎉该项目核心代码已经并入[FunASR](https://github.com/alibaba-damo-academy/FunASR) |
| 18 | + |
| 19 | +本仓库仅对模型做了转换，只采用ONNXRuntime推理引擎 |
| 20 | + |
| 21 | +## TODO |
17 | 22 | - [ ] 整合vad + asr + pun三个模型，打造可部署使用的方案 |
18 | 23 |
|
19 | | -#### 使用步骤 |
20 | | -1. Install |
21 | | - 1. 安装`rapid_paraformer` |
22 | | - ```bash |
23 | | - pip install rapid_paraformer |
24 | | - ``` |
25 | | - 2. 下载**resources.zip** ([Google Drive](https://drive.google.com/drive/folders/1RVQtMe0eB_k6G5TJlmXwPELx4VtF2oCw?usp=sharing) | [百度网盘](https://pan.baidu.com/s/1zf8Ta6QxFHY3Z75fHNYKrQ?pwd=6ekq)) |
26 | | - ```bash |
27 | | - resources |
28 | | - ├── [ 700] config.yaml |
29 | | - └── [4.0K] models |
30 | | - ├── [ 11K] am.mvn |
31 | | - ├── [824M] asr_paraformerv2.onnx |
32 | | - └── [ 50K] token_list.pkl |
33 | | - ``` |
34 | | - 3. **asr_paraformerv2.onnx**文件可基于modescope下的notebook环境自助转换： |
35 | | - 1. 打开[快速体验](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) |
36 | | - 2. 打开notebook → Cell中输入以下命令, 执行即可。 |
37 | | - ```python |
38 | | - !python -m funasr.export.export_model --model-name 'damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch' --export-dir "./export" |
39 | | - ``` |
40 | | - |
41 | | -2. 使用 |
42 | | - ```python |
43 | | - from rapid_paraformer import RapidParaformer |
44 | | -
|
45 | | - config_path = "resources/config.yaml" |
46 | | -
|
47 | | - paraformer = RapidParaformer(config_path) |
48 | | -
|
49 | | - wav_path = [ |
50 | | - "test_wavs/0478_00017.wav", |
51 | | - "test_wavs/asr_example_zh.wav", |
52 | | - ] |
53 | | -
|
54 | | - result = paraformer(wav_path) |
55 | | - print(result) |
56 | | - ``` |
57 | | -3. 查看结果 |
58 | | - ```text |
59 | | - ['y', '欢迎大家来体验达摩院推出的语音识别模型'] |
| 24 | +## 安装 |
| 25 | +```bash |
| 26 | +pip install rapid_paraformer |
| 27 | +``` |
| 28 | + |
| 29 | +## 模型下载 |
| 30 | +方法一：从Hugging Face上下载（[link](https://huggingface.co/SWHL/RapidParaformer)） |
| 31 | +```python |
| 32 | +from rapid_paraformer import download_hf_model |
| 33 | + |
| 34 | +download_hf_model(repo_id="SWHL/RapidParaformer", save_dir=".") |
| 35 | +``` |
| 36 | + |
| 37 | +方法二：([Google Drive](https://drive.google.com/drive/folders/1RVQtMe0eB_k6G5TJlmXwPELx4VtF2oCw?usp=sharing) | [百度网盘](https://pan.baidu.com/s/1zf8Ta6QxFHY3Z75fHNYKrQ?pwd=6ekq)) |
| 38 | +```bash |
| 39 | +resources |
| 40 | +├── [ 700] config.yaml |
| 41 | +└── [4.0K] models |
| 42 | + ├── [ 11K] am.mvn |
| 43 | + ├── [824M] asr_paraformerv2.onnx |
| 44 | + └── [ 50K] token_list.pkl |
| 45 | +``` |
| 46 | + |
| 47 | +## 模型转换 |
| 48 | +基于modescope下的notebook环境自助转换： |
| 49 | +1. 打开[快速体验](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) |
| 50 | +2. 打开notebook → Cell中输入以下命令, 执行即可。 |
| 51 | + ```bash |
| 52 | + !python -m funasr.export.export_model --model-name 'damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch' --export-dir "./export" |
60 | 53 | ``` |
| 54 | + |
| 55 | +## 使用 |
| 56 | +```python |
| 57 | +from rapid_paraformer import RapidParaformer |
| 58 | +
|
| 59 | +config_path = "resources/config.yaml" |
| 60 | +
|
| 61 | +paraformer = RapidParaformer(config_path) |
| 62 | +
|
| 63 | +wav_path = [ |
| 64 | + "test_wavs/0478_00017.wav", |
| 65 | + "test_wavs/asr_example_zh.wav", |
| 66 | +] |
| 67 | +
|
| 68 | +result = paraformer(wav_path) |
| 69 | +print(result) |
| 70 | +# ['y', '欢迎大家来体验达摩院推出的语音识别模型'] |
| 71 | +``` |
0 commit comments