File size: 1,713 Bytes
9cda4dd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
84ddd29
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
# AliCTTransformerPunc 模型介绍
### 使用方法
下载模型
```shell
git clone https://github.com/manyeyes/AliCTTransformerPunc.git
cd /AliCTTransformerPunc/AliCTTransformerPunc.Examples
git clone https://huggingface.co./manyeyes/punc_ct-transformer_zh-cn-common-vocab272727-onnx.git
```
编辑 Program.cs 中的 modelName 值:
```csharp
string modelName="punc_ct-transformer_zh-cn-common-vocab272727-onnx";
//[fp32模型/int8模型],二选一
//fp32模型
//string modelFilePath = applicationBase + "./"+ modelName + "/model.onnx";
//int8模型
string modelFilePath = applicationBase + "./"+ modelName + "/model_quant.onnx";
```
将文件夹 punc_ct-transformer_zh-cn-common-vocab272727-onnx 中的文件属性-》复制到输出目录-》如果较新则复制。
**或者**
编辑 AliCTTransformerPunc.Examples.csproj ,添加
```xml
  <ItemGroup>
    <None Update="punc_ct-transformer_zh-cn-common-vocab272727-onnx\punc.yaml">
      <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
    </None>
    <None Update="punc_ct-transformer_zh-cn-common-vocab272727-onnx\tokens.txt">
      <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
    </None>
  </ItemGroup>
```

启动 AliCTTransformerPunc.Examples


## 通过modelscope了解更多
### 原模型
- pytorch
https://www.modelscope.cn/models/damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch/summary
- 基于FunASR进行模型的微调和推理:
FunASR框架支持魔搭社区开源的工业级的语音识别模型的training & finetuning,使得研究人员和开发者可以更加便捷的进行语音识别模型的研究和生产,目前已在Github开源:https://github.com/alibaba-damo-academy/FunASR 。