Releases: jianchang512/pyvideotrans
windows-videotrans-v3.25
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: 音量调整对所有渠道均生效，数值代表增减百分比
- Fix: stt语音识别渠道提示不支持的格式错误
- Fix: #639 #630 #636
- Fix: 克隆声音时对中文英文 文本规范化
- Fix: 试听配音报错提示
- Feat: 添加孟加拉语
- Fix: 单个视频翻译中途停止时清空字幕区
- Fix: 默认选中语音降噪
- Feat: Gemini语音识别时增加vad切割为小片段再挨个发送请求识别，以避免时间戳不准确问题，翻译设置Gemini pro中启用(免费Gemini账号勿启用，会导致大量429错误)
Win v3.24 完整包下载3.5G(含tiny/medium模型)
如果未安装过旧版本，请在此下载完整版，如需cuda加速，需有英伟达显卡并且安装cuda12.x及cudnn9
百度网盘下载地址: https://pan.baidu.com/s/19c2HrofGQy3KSJayLWNBtw?pwd=cx5v
Huggingface(墙外地址): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-v3.24.7z?download=true
v3.25补丁包165MB
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
GitHub地址: https://github.com/jianchang512/pyvideotrans/releases/download/v3.19/win-PatchUpdate-3.25.7z
百度网盘下载地址: https://pan.baidu.com/s/1151v1KVXb1wnadTKHykAVA?pwd=bpvt
若补丁覆盖后打开出错，请升级显卡驱动、升级cuda到12.x、升级cudnn到cudnn9，并重新下载完整包，然后再使用该补丁包覆盖。
或下载该压缩包 https://pan.baidu.com/s/1LOjr9ampzqp7eBrjNgSTHQ?pwd=wrkn 解压后(也可从此地址下载
https://github.com/jianchang512/pyvideotrans/releases/download/v3.15/cuda-torch-pathupdate.7z )，复制其内的_internal文件夹覆盖软件目录内的_internal文件夹
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/v3.00/api.exe
所有模型下载地址
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
windows-videotrans-v3.18
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: 克隆语音时，短于4s的原音频复制延长一倍时长
- Feat: 添加语音降噪功能以提高转录精度，在线从modelscope.cn下载模型
- Fix: 高级选项添加音频补偿设置，用于抵消字幕对齐时随着时间增加配音逐渐落后画面
- Feat: 高级选项中增加 CUDA视频硬件解码
- Fix: 更新faster-whisper下 large-v3-turbo模型
- Feat: 单个视频翻译时，目标语言字幕区优化为可单独为每行设置角色
- Fix: 字幕文字翻译缓存移动至 软件目录的 /tmp下，不需要缓存时可删除
- Change: torch升级到2.5，cuda加速最低需12.x版本
Win v3.18 完整包下载3.4G(含tiny/medium模型)
如果未安装过旧版本，请在此下载完整版，如需cuda加速，需有英伟达显卡并且安装cuda12.x及cudnn9
百度网盘下载地址: https://pan.baidu.com/s/1fsEIHJvSaA3JyHAXyDTwxw?pwd=y2m7
Huggingface(墙外地址): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-v3.18.7z?download=true
v3.18补丁包92MB
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
GitHub地址: https://github.com/jianchang512/pyvideotrans/releases/download/v3.15/win-PatchUpdate-3.18.7z
百度网盘下载地址: https://pan.baidu.com/s/17u79I8eQOo9To3tZFj-ohA?pwd=gbm3
若补丁覆盖后打开出错，请升级显卡驱动、升级cuda到12.x、升级cudnn到cudnn9，并重新下载完整包，然后再使用该补丁包覆盖。
或下载该压缩包 (https://pan.baidu.com/s/1LOjr9ampzqp7eBrjNgSTHQ?pwd=wrkn) 解压后，复制其内的
_internal文件夹覆盖软件目录内的_internal文件夹
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/v3.00/api.exe
所有模型下载地址
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
windows-v3.14
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Feat: AI翻译时，并且选中了发送完整字幕，将启用三步反思法翻译，效果可能更好，但仅限在线大模型，不支持本地模型(显然效果会很差)
- Fix: 修复Deepl选择发音语言自动时bug
v3.11 完整包下载3.4G(含tiny/medium模型)
如果未安装过旧版本，请在此下载完整版
百度网盘: https://pan.baidu.com/s/1FTKHFM7l9ndbTmX0Z2uovg?pwd=9dm8
Huggingface(墙外): https://huggingface.co/spaces/mortimerme/s4/resolve/main/windows-videotrans-v3.11.7z?download=true
v3.14补丁包84MB
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
GitHub: https://github.com/jianchang512/pyvideotrans/releases/download/v3.00/win-PatchUpdate-3.14.7z
百度网盘: https://pan.baidu.com/s/1a-SlQZkDpT8gaXJs_IglOA?pwd=rvgu
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/v3.00/api.exe
所有模型下载地址
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
windows-v2.98
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: #599
- Feat:
zh_recong和SenseVoice识别渠道合成为FunASR中文 - Fix: 腾讯翻译、百度翻译、微软翻译 中文简繁语言代码修正
v2.98 完整包下载3.4G(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/1L8MhjsJ47Q_EVr9qKFdBag?pwd=r8j5
Huggingface(墙外): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-v2.98.7z?download=true
v2.99 补丁包84MB/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/v2.97/win-PatchUpdate-2.99.7z
百度网盘: https://pan.baidu.com/s/1DZhVpvvtyOnuE2JlKxTCAQ?pwd=y2uw
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!
windows-v2.96
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_namekeyError - 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!
windows-v2.91
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: F5-TTS 数字转为中文再朗读
- Feat: 配音渠道增加 F5-TTS
- Fix: 单个视频时双字幕区域显示
- Feat: 增加 SenseVoice 语音识别
- Feat: 配音渠道新增“字节火山语音合成” https://www.volcengine.com/docs/6561/79817
- Fix: #585 #586
- Fix: 修复 GPT-SoVITS api_v2 语速参数错误
- Feat: 添加文档教程
- Feat: 点击 “翻译渠道”和“配音渠道”时增加并发数控制和暂停时长
- Fix #579
v2.90 完整包下载3.2G(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
百度网盘: https://pan.baidu.com/s/12iRxMv7iAQI7mtwgNbtbqg?pwd=afeu
Huggingface(墙外): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-2.90.7z?download=true
v2.91 补丁包160MB/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/v2.85/win-PatchUpdate-2.91.7z
百度网盘: https://pan.baidu.com/s/1RRvrAMb_ag9wsIuUGJFt5w?pwd=b2ni
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!
windows-v2.84
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Feat: 点击 “翻译渠道”和“配音渠道”时增加并发数控制和暂停时长
- Fix: edgetTTS增加代理设置
- Feat: 处理单个视频时，增加双字幕区域，同时显示原始字幕和翻译字幕
- Feat: 优化界面设置，点击“语音识别”和“嵌入字幕”时，可进一步设置详细参数
- Fix: #573
- Fix: 当选择
faster-whisper本地时，点击左侧语音识别可详细设置语音切割时静音时长、阈值、最大最小语音片段等详细参数 - Fix: 修复特殊路径ffmpeg报错
v2.80 完整包下载3.2G(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
百度网盘: https://pan.baidu.com/s/1a0Nc41t_mkn_-C_saqqSXQ?pwd=wnbb
Huggingface(墙外): https://huggingface.co/spaces/mortimerme/sovits/resolve/main/win-videotrans-2.80.7z?download=true
v2.84 补丁包160MB/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/v2.79/win-PatchUpdate-2.84.7z
百度网盘: https://pan.baidu.com/s/155QRLIRlUBH6Oi9YSP4z6Q?pwd=dbtd
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!
windows-v2.77
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: 优化报错提示，移除qtpy依赖
- Fix: 优化重新断句策略，修正翻译后空白行导致时间错乱
- Fix: openai-whisper int str 比较bug
- Feat: Google 翻译渠道默认发送srt格式字符串
- Feat: 增加faster-whisper模式下large-v3-turbo模型支持
因faster-whisper官方尚不不支持turbo模型，源码部署需修改 faster_whisper包下 utils.py 文件第23行增加一行代码
"large-v3-turbo": "Systran/faster-whisper-large-v3-turbo",
v2.73 完整包下载3.2G(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
百度网盘: https://pan.baidu.com/s/1yahkCsWdYyuckPj1LWn1vg?pwd=fsfg
Huggingface(墙外): https://huggingface.co/spaces/mortimerme/s4/resolve/main/win-videotrans-2.73.7z?download=true
v2.78 补丁包160MB/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/v2.72/win-PatchUpdate-2.78.7z
百度网盘: https://pan.baidu.com/s/1VYR2MwAqaKplQfauQ3bFgw?pwd=scb2
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!
Windows-v2.71
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: 修复多角色配音
- Feat: 批量字幕翻译时AI渠道增加模型选择
- Fix: #554
- Feat: 音视频转字幕、批量翻译、批量配音记住上次选项
- Fix: 优化启动速度
- Fix: #554
- Fix: #553
- Feat: 增加 openai-whisper/large-v3-turbo
- 工具/选项-高级选项中增加“faster/openai-whisper识别后重新断句”选项，默认重新划分句子
v2.70 完整包下载/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
百度网盘: https://pan.baidu.com/s/1_r-at_r_K3usku2Yj4lRxw?pwd=ziqc
GitHub: https://github.com/jianchang512/pyvideotrans/releases/download/v2.66/win-videotrans-2.70.7z
v2.71 补丁包/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/v2.66/win-PatchUpdate-2.71.7z
百度网盘: https://pan.baidu.com/s/1swWlcZL2avEdjYbZeHdANQ?pwd=7iad
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!
Windows-v2.65
预打包版仅适用于Windows10/11，MacOS和Linux请源码部署
- Fix: windows上无法创建以
.或 - 工具/选项-高级选项中增加“faster/openai-whisper识别后重新断句”选项，默认重新划分句子
v2.65 完整包下载/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
百度网盘: https://pan.baidu.com/s/1theP_xdrFgVqXAg4cF68Rw?pwd=psf1
GitHub: https://github.com/jianchang512/pyvideotrans/releases/download/v2.64/win-videotrans-2.65.7z
v2.65 补丁包/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/v2.64/win-PatchUpdate-2.65.7z
百度网盘: https://pan.baidu.com/s/1XKvAZ3I1595-tqykKyLpwQ?pwd=3rei
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!