v2.92

> 预打包版仅适用于Windows10/11，MacOS和Linux请源码部署


- Fix: 修复识别为空时index out of range 错误
- Fix: #596 
- Fix: `target_language_name` keyError
- Fix: `yarl` 版本更新
- Feat: 新增 deepgram.com 语音识别api
- Feat: 新增中文重新断句功能，使用阿里中文标点恢复模型 ct-punc-c
- Fix: 字幕翻译返回结果错误
- Fix: AI翻译时 prompt 增加示例
- Fix: #591 


## v2.94 完整包下载3.6G(含tiny/medium模型)/complete package

> 如果未安装过旧版本，请在此下载完整版/If have not installed an older version, download the full version here

**解压后 videotrans/cfg.json和videotrans/params.json 这2个文件请删掉，否则报错**

百度网盘:  https://pan.baidu.com/s/1zUUw3LptF62kJB7ETFGCtg?pwd=hbyh

Huggingface(墙外): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-v2.94.7z?download=true

 

## v2.96 补丁包360MB/patchUpdate  package

> 如果已安装过2.x版本，可下载补丁包后解压在sp.exe所在目录，覆盖已有sp.exe和文件夹
> If have already installed version 2.x, can download the patch package and unzip it in the sp.exe directory, overwriting the existing

GitHub: https://github.com/jianchang512/pyvideotrans/releases/download/v2.92/win-PatchUpdate-2.96.7z

百度网盘: https://pan.baidu.com/s/1QMFz636J8GFaiJXvUO-flA?pwd=kr6f



## v2.92+ 升级到 v2.96补丁包 50M

下载(仅限2.92+->2.95):  https://github.com/jianchang512/pyvideotrans/releases/download/v2.92/2.92-to-2.96.7z

----

> api.exe 下载，api.exe是用于通过http请求调用的文件，不可独立使用，需先下载完整包，然后下载api.exe和 sp.exe放在同一目录下
> 
> api.exe下载地址：https://github.com/jianchang512/pyvideotrans/releases/download/v2.92/api.exe


----

### 所有模型下载地址/Download all models

> 为避免压缩包体积过大，预打包版只内置最小模型 tiny，识别效果不佳，效果更好的模型请点击下载  
> In order to avoid the compressed package volume is too large, the pre-packaged version is only built-in the smallest model tiny, the recognition effect is not good, the effect of better models please click to download!  

https://github.com/jianchang512/stt/releases/tag/0.0