#4656 TaskCfgVTT(is_cuda=True, uuid='52ea7523e3', cache_folder='D:/videotrans/tmp/6784/52ea7523e3', target_dir='D:/Video/_vide

120.229* Posted at: 2 hours ago 👁8

语音识别阶段出错 [openai-whisper(本地)] 出错了,可能内存或显存不足 Model:medium.en GPU0
Traceback (most recent call last):
File "videotrans\configure\_base.py", line 281, in _new_process
File "videotrans\process\signelobj.py", line 80, in submit_task_gpu
File "concurrent\futures\process.py", line 720, in submit
concurrent.futures.process.BrokenProcessPool: A child process terminated abruptly, the process pool is not usable anymore

Traceback (most recent call last):
File "videotrans\configure\_base.py", line 281, in _new_process
File "videotrans\process\signelobj.py", line 80, in submit_task_gpu
File "concurrent\futures\process.py", line 720, in submit
concurrent.futures.process.BrokenProcessPool: A child process terminated abruptly, the process pool is not usable anymore

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "videotrans\task\job.py", line 105, in run
File "videotrans\task\trans_create.py", line 360, in recogn
File "videotrans\recognition\__init__.py", line 292, in run
File "videotrans\recognition\_base.py", line 143, in run
File "videotrans\recognition\_overall.py", line 31, in _exec
File "videotrans\recognition\_overall.py", line 73, in _openai
File "videotrans\configure\_base.py", line 303, in _new_process
RuntimeError: 出错了,可能内存或显存不足 Model:medium.en GPU0
Traceback (most recent call last):
File "videotrans\configure\_base.py", line 281, in _new_process
File "videotrans\process\signelobj.py", line 80, in submit_task_gpu
File "concurrent\futures\process.py", line 720, in submit
concurrent.futures.process.BrokenProcessPool: A child process terminated abruptly, the process pool is not usable anymore
TaskCfgVTT(is_cuda=True, uuid='52ea7523e3', cache_folder='D:/videotrans/tmp/6784/52ea7523e3', target_dir='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4', source_language='英语', source_language_code='en', source_sub='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4/en.srt', source_wav='D:/videotrans/tmp/6784/52ea7523e3/en.wav', source_wav_output='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4/en.m4a', target_language='简体中文', target_language_code='zh-cn', target_sub='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4/zh-cn.srt', target_wav='D:/videotrans/tmp/6784/52ea7523e3/target.wav', target_wav_output='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4/zh-cn.m4a', name='D:/Video/08C - 蜡烛 设置和信号K线.mp4', noextname='08C - 蜡烛 设置和信号K线', basename='08C - 蜡烛 设置和信号K线.mp4', ext='mp4', dirname='D:/Video', shound_del_name=None, translate_type=0, tts_type=0, volume='+0%', pitch='+0Hz', voice_rate='+0%', voice_role='No', voice_autorate=False, video_autorate=False, remove_silent_mid=False, align_sub_audio=True, detect_language='en', recogn_type=1, model_name='medium.en', shibie_audio=None, remove_noise=False, enable_diariz=False, nums_diariz=0, rephrase=0, fix_punc=False, subtitle_language=None, app_mode='tiqu', subtitles='', targetdir_mp4='D:/Video/_video_out/08C - 蜡烛 设置和信号K线-mp4/08C - 蜡烛 设置和信号K线.mp4', novoice_mp4='D:/videotrans/tmp/6784/52ea7523e3/novoice.mp4', is_separate=True, embed_bgm=True, instrument='D:/videotrans/tmp/6784/52ea7523e3/instrument.wav', vocal='D:/videotrans/tmp/6784/52ea7523e3/vocal.wav', back_audio='', clear_cache=False, background_music=None, subtitle_type=0, only_out_mp4=False, recogn2pass=True, output_srt=0, copysrt_rawvideo=False)


=
system:Windows-10-10.0.22631-SP0
version:v3.98
frozen:True
language:zh
root_dir:D:/videotrans

1 Replies

这是由于显存不足导致子进程崩溃。medium.en 模型在 GPU 上运行时需要较多显存,而你同时启用了 二次识别recogn2pass=True),这会在已有任务基础上再额外占用显存,更容易耗尽资源。

解决方案(按推荐顺序尝试):

  1. 取消“二次识别”:在主界面取消勾选 二次识别 选项,避免同时运行两轮语音识别。
  2. 更换更小的模型:将语音识别模型从 medium.en 改为 small.enbase.en,可大幅降低显存占用。
  3. 调整高级选项:进入菜单栏 工具/选项高级选项语音识别参数,进行如下修改:

    • CUDA数据类型:改为 float16int8
    • beam_size:改为 1
    • best_of:改为 1
    • 上下文:取消勾选(设为 false
  4. 关闭其他占用显存的程序:确保没有同时运行其他 GPU 应用(如游戏、其他识别任务)。

请查阅相关文档:

Post Your Reply
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.

Related resource