Page 51 BBS.pyVideoTrans.com TaskCfgSTT(is_cuda=True, uuid='90e5c416ef', cache_folder='E:/win-pyvideotrans-v3.99-428/tmp/15200/90e5c416ef', target_dir='E:/_pyVideotrans Output/recogn', source_language=None, source_language_code=None, source_sub=None, source_wav=None, source_wav_output=None, target_language=None, target_language_code=None, target_sub='E:/_pyVideotrans Output/recogn/4. Operators.srt', target_wav=None, target_wav_output=None, name='E:/0DayDown/教程_Tutorials/Unreal/Verse Bootcamp UEFN Academy/2 - Core Basics/4. Operators.mp4', noextname='4. Operators', basename='4. Operators.mp4', ext='mp4', dirname='E:/0DayDown/教程_Tutorials/Unreal/Verse Bootcamp UEFN Academy/2 - Core Basics', shound_del_name=None, detect_language='en', recogn_type=0, model_name='large-v3-turbo', shibie_audio='E:/win-pyvideotrans-v3.99-428/tmp/15200/90e5c416ef/4. Operators-1777717918.1931195.wav', remove_noise=False, enable_diariz=False, nums_diariz=0, rephrase=1, fix_punc=True)

#4383 TaskCfgSTT(is_cuda=True, uuid='90e5c416ef', cache_folder='E:/win-pyvideotrans-v3.99-428/tmp/15200/90e5c416ef', target_dir='E:/_pyVideotrans Output/recogn', source_language=None, so...

[Last reply:1 month ago] 由于输入视频文件路径中含有中文和空格，导致 FFmpeg 命令行在创建输出文件时失败（Invalid argument）。这在 Windows 系统中尤为常见，即使路径被引号包围也可能触发编码兼容性问题。解决方案将原始视频 E:/0DayDown/教程_Tutorials/Unreal/Verse Bootcamp UEFN Academy/2 - Core Basics/4. Operators...

103.172* Posted at: 1 month ago

#4382 submitting packet to decoder: Invalid data found when processing input submitting packet to decoder: Invalid data found when processing input submitting packet to decoder: Invalid ...

💬 1

[Last reply:1 month ago] 这是 FFmpeg 解码输入文件时遇到损坏或无法识别的数据所致。错误信息 Invalid data found when processing input 表明视频文件 A:/kasi1111/01.mp4 可能已损坏、编码异常，或存放在不稳定的磁盘（如网络驱动器、移动硬盘）上导致读取数据错误。请按以下步骤排查：将视频复制到本地磁盘（如 D 盘根目录），并将文件名改为简短的英文/数字，例如 D:...

91.243* Posted at: 1 month ago

#4381 openai.NotFoundError: Error code: 404 - {'detail': 'Not Found'}...

💬 1

[Last reply:1 month ago] This is caused by an incorrect API endpoint or model name for the OpenAI TTS service. The 404 error indicates the requested resource was not found on the server.Verify that you have correctly filled i...

2405:4802* Posted at: 1 month ago

#4380 videotrans.configure._except.StopRetry: Could not fetch config for http://127.0.0.1:8880/v1/audio/speech/...

💬 1

[Last reply:1 month ago] 错误源于 OmniVoice 服务未正确启动或 API 地址配置有误，导致 pyVideoTrans 无法获取 Gradio 服务的配置文件。解决方案：确认 OmniVoice 服务已在终端中启动，并保持运行状态，没有报错。检查 pyVideoTrans 中 OmniVoice 的 API 地址设置：菜单 -> TTS设置 -> F5-TTS界面（OmniVoice 复用此设置）应填写...

119.39* Posted at: 1 month ago

#4379 TaskCfgTTS(is_cuda=False, uuid='095993aa84', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/5965/095993aa84', target_dir='/Users/junyouchen/Downloads', source_language=No...

💬 1

[Last reply:1 month ago] 这是 pyVideoTrans 中 piper TTS(本地) 模块的内部代码缺陷（UnboundLocalError），变量 e 在异常捕获前未被赋值就直接使用。解决方案：临时绕开此错误：在主界面更换配音渠道，推荐使用 Edge-TTS（免费且稳定），或部署其他本地 TTS 如 CosyVoice、GPT-SoVITS 等。升级到最新代码：你是 macOS 源码部署的，请在项目目录执行：git ...

51.75* Posted at: 1 month ago

#4378 TaskCfgTTS(is_cuda=False, uuid='77de970a7c', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/5965/77de970a7c', target_dir='/Users/junyouchen/Downloads', source_language=No...

💬 1

[Last reply:1 month ago] 这是由 Piper TTS 内部代码缺陷，以及文件路径含有中文特殊符号共同导致的错误。核心原因：UnboundLocalError: local variable 'e' referenced before assignment 是 Piper TTS 模块的一个代码 Bug。程序企图在出错时引用错误变量，但因为文件路径问题，程序在执行初期就陷入异常，导致这个 Bug 暴露了出...

51.75* Posted at: 1 month ago

#4377 TaskCfgVTT(is_cuda=True, uuid='11339c2d22', cache_folder='C:/398/tmp/18108/11339c2d22', target_dir='C:/Users/trise/Downloads/_video_out/0428-mp4', source_language='Simplified Chine...

💬 1

[Last reply:1 month ago] The error is due to insufficient GPU video memory when using the medium model with CUDA enabled, made worse by the enabled "二次识别" (recogn2pass) option which doubles memory usage.Disable "二次识别" On the...

84.17* Posted at: 1 month ago

#4376 LLM重新断句的结果，主要受哪些参数设置的影响...

💬 7

[Last reply:1 month ago] 字幕对齐主要在视频翻译的最后合成阶段执行，即在配音完成后、最终视频输出前。具体过程包括以下环节：配音加速 / 视频慢速：当配音时长与原画不对应时，软件根据主界面的勾选项自动加速配音或放慢视频，强制匹配原片段时长。对齐字幕声音：若不使用配音加速或视频慢速，可勾选此项强制对齐字幕与配音的时间轴。二次识别：若勾选，软件会在配音完成后对整个配音音频再次进行语音识别，生成时间轴精准匹配配音的新字幕，确保字幕...

103.172* Posted at: 1 month ago

#4375 openai.LengthFinishReasonError: Could not parse response content as the length limit was reached - CompletionUsage(completion_tokens=916, prompt_tokens=3180, total_tokens=4096, com...

💬 1

[Last reply:1 month ago] Translation failed because the AI model reached its token limit. The prompt (3180 tokens) plus the allowed completion length caused a 4096 total token cap to be exceeded, truncating the output and mak...

2405:4802* Posted at: 1 month ago

#4374 dubbing speedup esta deixando a voz robotica...

💬 1

[Last reply:1 month ago] The robotic voice occurs because the "Dubbing Speedup" option forces synthesized audio to accelerate beyond natural limits, causing distortion. This typically happens when translated speech is signifi...

191.6* Posted at: 1 month ago

#4373 TaskCfgVTT(is_cuda=False, uuid='3f8c4968c8', cache_folder='C:/Users/HAMZA/Desktop/doblag youtube/win-pyvideotrans-v3.99-428/tmp/6932/3f8c4968c8', target_dir='doblaj 1-mp4', source_...

💬 1

[Last reply:1 month ago] This is caused by the space in your video file path. The subprocess call to ffprobe fails because the path C:/Users/HAMZA/Desktop/capcut vidio/doblaj 1.mp4 contains spaces and the command is not prope...

105.159* Posted at: 1 month ago

#4372 TaskCfgVTT(is_cuda=True, uuid='617e6c9a85', cache_folder='D:/win-pyvideotrans-v3.99-428/tmp/2152/617e6c9a85', target_dir='G:/John Templeton/_video_out/John Templeton On Finding Tru...

💬 1

[Last reply:1 month ago] 该错误是由于 CUDA 显卡被其他程序占用、显存不足或当前 GPU 状态不可用导致的。你的配置中开启了 CUDA 加速并选用了 large-v3 模型（需 10GB+ 显存），且启用了“二次识别”（recogn2pass=True），这会进一步推高显存占用。解决方案（按顺序尝试）：关闭“二次识别”：在主界面右上角取消勾选“二次识别”复选框。该功能会在配音后再次调用语音识别，容易与首次识别争抢显存。...

124.77* Posted at: 1 month ago

#4371 Warning: flash-attn is not installed. Will only run the manual PyTorch version. Please install flash-attn for faster inf......

💬 5

[Last reply:1 month ago] flash-attn 警告是因为此库依赖 CUDA，在 macOS 上几乎无法正常安装（无 NVIDIA GPU 时也不会生效），可以安全忽略，不会阻止程序运行。sox: command not found 则是缺少音频工具，必须安装。解决步骤：安装 SoX：brew install sox忽略 Flash Attention 警告，或在虚拟环境中尝试安装（大概率失败）：使用项目自带的 uv 环境...

51.75* Posted at: 1 month ago

#4370 TaskCfgTTS(is_cuda=False, uuid='66c21c82f1', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/66102/66c21c82f1', target_dir='/Users/junyouchen/Downloads', source_language=N...

💬 1

[Last reply:1 month ago] 这是由于未安装 qwen_tts 模块所致，Qwen3-TTS(本地) 渠道需要该第三方库。解决方案：在终端中进入软件源码目录 /Users/junyouchen/Desktop/pyvideotrans；使用 uv 添加依赖：uv add qwen_tts或使用 pip 安装（不推荐，可能干扰 uv 环境）：uv run pip install qwen_tts安装完成后重新启动软件（uv ru...

119.39* Posted at: 1 month ago

#4369 为pyVideoTrans配置cuda环境一定要安装12.8版本吗？目前最新的cudatoolkit已经更新到13.2了，是否安装最新版本更好？还是必须安装12.8这一特定版本...

💬 3

[Last reply:1 month ago] 是的，可以安装最新版 cuDNN（例如 9.21），只要版本 ≥ 9.11 即满足要求。安装时请务必下载针对 CUDA 12.8 编译的 cuDNN 版本（NVIDIA 官网会提供对应选项），否则可能出现库不匹配。安装后同样需将文件复制到 CUDA 12.8 的 bin、include、lib 目录。请查阅相关文档：Windows 上安装 CUDA 12.8 和 cuDNN 9.11已安...

103.3* Posted at: 1 month ago

#4368 恢复标点能否不要恢复半句话的标点，也就是存在断句错误的句子，一句完整的话被强行切割到两条不同的字幕中，这种情况下，前半句一般会出现在字幕的末尾，这种不要加标点...

💬 1

[Last reply:1 month ago] 暂无法确定具体原因。恢复标点功能在内部逻辑上尚未提供针对“断句错误”场景的半句区分能力，知识库中亦无相关配置参数可供调整。建议提供您使用的软件版本、原始字幕片段示例以及期望的处理行为，以便开发者评估优化。若需立即解决，可尝试以下临时替代方案：取消勾选“恢复标点”，使用更准确的语音识别模型（如 large-v3）以减少断句错误。在翻译/合成前，手动编辑识别出的 SRT 字幕，将明显被错切的句子合并并...

2409:8a55* Posted at: 1 month ago

#4367 Traceback (most recent call last):File "/Users/junyouchen/Desktop! pyvideotrans/.venv/lib/python3.10/site-packages/sound......

💬 1

[Last reply:1 month ago] 这是因为系统中缺少 libsndfile 音频处理库导致 Python 的 soundfile 模块无法加载。解决方案：打开终端，使用 Homebrew 安装 libsndfile：brew install libsndfile若尚未安装 FFmpeg（软件必需），也一并安装：brew install ffmpeg安装完成后，重新运行软件。请注意：macOS 上源码部署还需确保 Python 3....

51.75* Posted at: 1 month ago

#4366 ValueError: The elements of the batch contain different keys. Cannot batch them ({'is_last', 'input_features', 'attention_mask'} != {'num_frames', 'is_last', 'input_features', 'att...

💬 1

[Last reply:1 month ago] 这是由于音频片段长度不一致导致 transformers 管道批处理时特征键不匹配（input_features 形状不同）。解决方案：在菜单栏工具/选项 → 高级选项 → 语音识别参数中，勾选 “Whisper预分割音频”，强制模型逐句处理，避免不统一长度的片段混合批处理。将 “最长语音持续秒数” 设为较小的值（如 6 秒），避免生成极长片段；“最短语音持续毫秒” 设为 1000ms 以上，...

2408:8262* Posted at: 1 month ago

#4365 TaskCfgVTT(is_cuda=True, uuid='7fe7d62dc3', cache_folder='D:/AI/pyvideotrans/tmp/167832/7fe7d62dc3', target_dir='D:/AI/pyvideotrans/video goc/_video_out/videoplayback-mp4', source_...

💬 1

[Last reply:1 month ago] This error is caused by an incompatibility between the cuBLAS library and your NVIDIA graphics card, most commonly seen on RTX 50-series GPUs or when a mixed CUDA/cuDNN version is used.Check your CUDA...

14.179* Posted at: 1 month ago

#4364 为字幕配音，拖入srt文件字幕配好音后，点击播放配音，为什么带着字幕里时间序号一起播放念啊？如何设置为字幕配音只让他念内容不念字幕里的时间序号？...

💬 1

[Last reply:1 month ago] 这是因为软件未能正确解析 SRT 字幕文件，将序号和时间轴也当作了待配音的文本，导致 TTS 朗读了这些内容。通常是由于你使用的软件版本存在该 bug，或导入的不是标准 SRT 但软件未能过滤。解决方法：更新主程序：下载最新的 sp.exe 补丁文件覆盖到安装目录（v3.99 已修复相关 bug）。下载地址：https://github.com/jianchang512/stt/release...

121.27* Posted at: 1 month ago

#4383 TaskCfgSTT(is_cuda=True, uuid='90e5c416ef', cache_folder='E:/win-pyvideotrans-v3.99-428/tmp/15200/90e5c416ef', target_dir='E:/_pyVideotrans Output/recogn', source_language=None, so...

#4382 submitting packet to decoder: Invalid data found when processing input submitting packet to decoder: Invalid data found when processing input submitting packet to decoder: Invalid ...

#4381 openai.NotFoundError: Error code: 404 - {'detail': 'Not Found'}...

#4380 videotrans.configure._except.StopRetry: Could not fetch config for http://127.0.0.1:8880/v1/audio/speech/...

#4379 TaskCfgTTS(is_cuda=False, uuid='095993aa84', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/5965/095993aa84', target_dir='/Users/junyouchen/Downloads', source_language=No...

#4378 TaskCfgTTS(is_cuda=False, uuid='77de970a7c', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/5965/77de970a7c', target_dir='/Users/junyouchen/Downloads', source_language=No...

#4377 TaskCfgVTT(is_cuda=True, uuid='11339c2d22', cache_folder='C:/398/tmp/18108/11339c2d22', target_dir='C:/Users/trise/Downloads/_video_out/0428-mp4', source_language='Simplified Chine...

#4376 LLM重新断句的结果，主要受哪些参数设置的影响...

#4375 openai.LengthFinishReasonError: Could not parse response content as the length limit was reached - CompletionUsage(completion_tokens=916, prompt_tokens=3180, total_tokens=4096, com...

#4374 dubbing speedup esta deixando a voz robotica...

#4373 TaskCfgVTT(is_cuda=False, uuid='3f8c4968c8', cache_folder='C:/Users/HAMZA/Desktop/doblag youtube/win-pyvideotrans-v3.99-428/tmp/6932/3f8c4968c8', target_dir='doblaj 1-mp4', source_...

#4372 TaskCfgVTT(is_cuda=True, uuid='617e6c9a85', cache_folder='D:/win-pyvideotrans-v3.99-428/tmp/2152/617e6c9a85', target_dir='G:/John Templeton/_video_out/John Templeton On Finding Tru...

#4371 Warning: flash-attn is not installed. Will only run the manual PyTorch version. Please install flash-attn for faster inf......

#4370 TaskCfgTTS(is_cuda=False, uuid='66c21c82f1', cache_folder='/Users/junyouchen/Desktop/pyvideotrans/tmp/66102/66c21c82f1', target_dir='/Users/junyouchen/Downloads', source_language=N...

#4369 为pyVideoTrans配置cuda环境一定要安装12.8版本吗？目前最新的cudatoolkit已经更新到13.2了，是否安装最新版本更好？还是必须安装12.8这一特定版本...

#4368 恢复标点能否不要恢复半句话的标点，也就是存在断句错误的句子，一句完整的话被强行切割到两条不同的字幕中，这种情况下，前半句一般会出现在字幕的末尾，这种不要加标点...

#4367 Traceback (most recent call last):File "/Users/junyouchen/Desktop! pyvideotrans/.venv/lib/python3.10/site-packages/sound......

#4366 ValueError: The elements of the batch contain different keys. Cannot batch them ({'is_last', 'input_features', 'attention_mask'} != {'num_frames', 'is_last', 'input_features', 'att...

#4365 TaskCfgVTT(is_cuda=True, uuid='7fe7d62dc3', cache_folder='D:/AI/pyvideotrans/tmp/167832/7fe7d62dc3', target_dir='D:/AI/pyvideotrans/video goc/_video_out/videoplayback-mp4', source_...

#4364 为字幕配音，拖入srt文件字幕配好音后，点击播放配音，为什么带着字幕里时间序号一起播放念啊？如何设置为字幕配音只让他念内容不念字幕里的时间序号？...

Related resource