README.md 2.43 KB
Newer Older
SWHL's avatar
SWHL committed
1
## Rapid ASR
2
3
<p align="left">
    <a href=""><img src="https://img.shields.io/badge/OS-Linux%2C%20Win%2C%20Mac-pink.svg"></a>
SWHL's avatar
SWHL committed
4
5
    <a href=""><img src="https://img.shields.io/badge/Python->=3.7,<=3.10-aff.svg"></a>
    <a href=""><img src="https://img.shields.io/badge/C++-aff.svg"></a>
6
7
</p>

SWHL's avatar
SWHL committed
8
- 🎉 推出知识星球[RapidAI私享群](https://t.zsxq.com/0duLBZczw),这里的提问会优先得到回答和支持,也会享受到RapidAI组织后续持续优质的服务。欢迎大家的加入。
SWHL's avatar
SWHL committed
9
10
- Paraformer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)
- 本仓库仅对模型做了转换,只采用ONNXRuntime推理引擎。该项目核心代码已经并入[FunASR](https://github.com/alibaba-damo-academy/FunASR)
SWHL's avatar
SWHL committed
11
- 项目仍会持续更新,欢迎关注。
SWHL's avatar
SWHL committed
12
- QQ群号:645751008
13

SWHL's avatar
SWHL committed
14
15
16
17
18
19
20
21
22
23
24
#### 📖文档导航
- 语音识别:
    - rapid_paraformer:
        - [rapid_paraformer-Python](./python/README.md)
        - [rapid_C++/C](./cpp_onnx/readme.md)
    - [rapid_wenet](https://github.com/RapidAI/RapidASR/tree/rapid_wenet)
        - [Python](https://github.com/RapidAI/RapidASR/tree/rapid_wenet/python)
        - [C++](https://github.com/RapidAI/RapidASR/tree/rapid_wenet/cpp)
    - [rapid_paddlespeech-Python](https://github.com/RapidAI/RapidASR/tree/rapid_paddlespeech)
- 标点符号
    - [RapidPunc](https://github.com/RapidAI/RapidPunc)
SWHL's avatar
SWHL committed
25

SWHL's avatar
SWHL committed
26
27
#### 📆TODO以及任务认领
- 参见这里:[link](https://github.com/RapidAI/RapidASR/issues/15)
SWHL's avatar
SWHL committed
28
29
30
31
32
33
34
35
36

#### 🎨整体框架
```mermaid
flowchart LR

A([wav]) --RapidVad--> B([各个小段的音频]) --RapidASR--> C([识别的文本内容]) --RapidPunc--> D([最终识别内容])
```

#### 📣更新日志
SWHL's avatar
SWHL committed
37
38
<details>
<summary>详情</summary>
SWHL's avatar
SWHL committed
39

SWHL's avatar
SWHL committed
40
41
- 2023-02-25
   - 添加C++版本推理,使用onnxruntime引擎,预/后处理代码来自: [FastASR](https://github.com/chenkui164/FastASR)
SWHL's avatar
SWHL committed
42
43
44
45
46
47
48
49
50
- 2023-02-14 v2.0.3 update:
  - 修复librosa读取wav文件错误
  - 修复fbank与torch下fbank提取结果不一致bug
- 2023-02-11 v2.0.2 update:
  - 模型和推理代码解耦(`rapid_paraformer``resources`
  - 支持批量推理(通过`resources/config.yaml``batch_size`指定)
  - 增加多种输入方式(`Union[str, np.ndarray, List[str]]`
- 2023-02-10 v2.0.1 update:
  - 添加对输入音频为噪音或者静音的文件推理结果捕捉。
SWHL's avatar
SWHL committed
51

SWHL's avatar
SWHL committed
52
</details>