#8847 语音识别阶段出错[faster-whisper(内置)] File not found: F:/新建文件夹 (3)/tmp/28488/7c274c4b2e/1737641048710940004-1795054181.8584142.wa

108.181**7 Posted at: 3 hours ago

语音识别阶段出错[faster-whisper(内置)] File not found: F:/新建文件夹 (3)/tmp/28488/7c274c4b2e/1737641048710940004-1795054181.8584142.wav
Traceback (most recent call last):

File "videotrans\task\job.py", line 35, in run

File "videotrans\task\job.py", line 98, in process_task

File "videotrans\task\speech2text.py", line 142, in recogn

File "videotrans\recognition\__init__.py", line 178, in run

File "videotrans\recognition\_base.py", line 101, in run

File "videotrans\recognition\_whisper.py", line 35, in _exec

File "pydub\audio_segment.py", line 808, in from_wav

File "pydub\audio_segment.py", line 651, in from_file

File "pydub\utils.py", line 60, in _fd_or_path_or_tempfile

FileNotFoundError: [Errno 2] No such file or directory: 'F:/新建文件夹 (3)/tmp/28488/7c274c4b2e/1737641048710940004-1795054181.8584142.wav'
trk=语音转录
self.cfg.name='C:/Users/Administrator/Desktop/可发视频/1737641048710940004.mp3',
self.cfg.target_dir='F:/新建文件夹 (3)/output',
self.cfg.cache_folder='F:/新建文件夹 (3)/tmp/28488/7c274c4b2e'
self.cfg.is_cuda=True
语音识别:faster-whisper(内置), large-v3-turbo, 发音语言: pt
最长语音持续(秒): 5s
最短语音持续(毫秒): 3000ms
静音分割持续毫秒: 600ms
语音阈值: 0.45
选择VAD: silero
计算数据类型:default
识别准确度beam_size:5
识别准确度best_of:5
启用上下文感知:False
重复惩罚:1.0
文本压缩率:2.4
输出格式: txt
中日韩字幕单行字符数:15, 其他语言字幕单行字符数:40
self.cfg=TaskCfgSTT(uuid='7c274c4b2e', name='C:/Users/Administrator/Desktop/可发视频/1737641048710940004.mp3', dirname='C:/Users/Administrator/Desktop/可发视频', noextname='1737641048710940004', basename='1737641048710940004.mp3', ext='mp3', target_dir='F:/新建文件夹 (3)/output', cache_folder='F:/新建文件夹 (3)/tmp/28488/7c274c4b2e', is_cuda=True, source_language='', source_language_code='', source_sub=None, source_wav=None, source_wav_output=None, target_language='', target_language_code='', target_sub='F:/新建文件夹 (3)/output/1737641048710940004.srt', target_wav=None, target_wav_output=None, detect_language='pt', recogn_type=0, model_name='large-v3-turbo', shibie_audio='F:/新建文件夹 (3)/tmp/28488/7c274c4b2e/1737641048710940004-1795054181.8584142.wav', remove_noise=False, enable_diariz=False, nums_diariz=0, rephrase=False, fix_punc=0)


=
system:Windows-10-10.0.26300-SP0
version:v4.19
frozen:True
language:zh_CN
root_dir:F:/新建文件夹 (3)

1 Replies

#1 DeepSeek reply at 3 hours ago

这是临时切片 wav 文件在写入或读取时丢失导致的,最常见的原因是软件所在路径 F:/新建文件夹 (3) 含有中文、空格和括号,以及原始文件名/目录过长过深,导致 ffmpeg 切分音频写入失败或被杀软拦截删除。

排查与解决步骤:

  1. 移动软件到纯英文短路径:把整个软件文件夹移动到 D:/pyvideotrans 或 F:/pvt 这类只含英文数字、无空格无括号的浅层目录,重新启动 sp.exe。
  2. 规范源文件路径:把 C:/Users/Administrator/Desktop/可发视频/1737641048710940004.mp3 复制到 D:/videos/ 下并改名为 1.mp3,从软件中重新选择该文件。
  3. 检查 F 盘剩余空间:tmp 目录位于 F 盘,若 F 盘空间不足(或为只读/无写权限),切割出的 wav 会写入失败。确保 F 盘有数 GB 以上空闲。
  4. 关闭安全软件:将软件目录整体加入杀毒软件/Windows Defender 的信任白名单,并以管理员权限运行;部分杀软会静默删除 tmp 内的临时 wav。
  5. 重试前勾选 清理已生成:清掉上次残留的损坏缓存文件,再重新执行识别。
  6. 若以上处理后仍报同一错误,请确认是否同时开启了多个 sp.exe 实例(共用 tmp 目录,任一实例退出会清空 tmp 导致其他实例报“文件不存在”);如有,请关闭其余实例后重试。

补充:该报错与模型、CUDA、识别参数无关,纯属临时文件路径/权限/磁盘问题,无需更换模型或调低参数。

请查阅相关文档:

Thinking...
Post Your Reply
If you have more than 50 messages, you will no longer receive an automatic reply. Please ask a new question!
Open source and free maintenance is not easy. If this project is helpful to you, please consider making a small donation to help the project continue to maintain and update.

Dondate via ko-fi.com

Related resource