#4609 TaskCfgVTT(is_cuda=False, uuid='fa11f1b02f', cache_folder='F://tmp/29700/fa11f1b02f', target_dir='F:/新建文件夹/1-mp4', sourc

112.250* Posted at: 1 day ago 👁20

语音识别阶段出错 [faster-whisper(本地)] Traceback (most recent call last):
File "videotrans\process\stt_fun.py", line 356, in faster_whisper
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

Traceback (most recent call last):
File "videotrans\task\job.py", line 105, in run
File "videotrans\task\trans_create.py", line 390, in recogn
File "videotrans\recognition\__init__.py", line 293, in run
File "videotrans\recognition\_base.py", line 143, in run
File "videotrans\recognition\_overall.py", line 33, in _exec
File "videotrans\recognition\_overall.py", line 106, in _faster
File "videotrans\configure\_base.py", line 289, in _new_process
RuntimeError: Traceback (most recent call last):
File "videotrans\process\stt_fun.py", line 356, in faster_whisper
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory
TaskCfgVTT(is_cuda=False, uuid='fa11f1b02f', cache_folder='F://tmp/29700/fa11f1b02f', target_dir='F:/新建文件夹/1-mp4', source_language='英语', source_language_code='en', source_sub='F:/新建文件夹/1-mp4/en.srt', source_wav='F://tmp/29700/fa11f1b02f/en.wav', source_wav_output='F:/新建文件夹/1-mp4/en.m4a', target_language='简体中文', target_language_code='zh-cn', target_sub='F:/新建文件夹/1-mp4/zh-cn.srt', target_wav='F://tmp/29700/fa11f1b02f/target.wav', target_wav_output='F:/新建文件夹/1-mp4/zh-cn.m4a', name='F:/Bonus - The Stephen Kelly trading method/1.mp4', noextname='1', basename='1.mp4', ext='mp4', dirname='F:/Bonus - The Stephen Kelly trading method', shound_del_name=None, translate_type=0, tts_type=0, volume='+0%', pitch='+0Hz', voice_rate='+0%', voice_role='No', voice_autorate=True, video_autorate=False, remove_silent_mid=False, align_sub_audio=True, detect_language='en', recogn_type=0, model_name='large-v3-turbo', shibie_audio=None, remove_noise=False, enable_diariz=False, nums_diariz=0, rephrase=0, fix_punc=False, subtitle_language=None, app_mode='biaozhun', subtitles='', targetdir_mp4='F:/新建文件夹/1-mp4/1.mp4', novoice_mp4='F://tmp/29700/fa11f1b02f/novoice.mp4', is_separate=False, embed_bgm=True, instrument=None, vocal='F://tmp/29700/fa11f1b02f/vocal.wav', back_audio='', clear_cache=True, background_music=None, subtitle_type=1, only_out_mp4=False, recogn2pass=True, output_srt=0, copysrt_rawvideo=False, loop_backaudio=1, backaudio_volume=0.8)


=
system:Windows-10-10.0.22631-SP0
version:v3.99
frozen:True
language:zh
root_dir:F:/

1 Replies

这是由于系统内存不足导致 faster-whisper 初始化时无法分配内存。您在日志中看到 is_cuda=False,说明未启用 GPU 加速,模型 large-v3-turbo 在 CPU 模式下需要大量物理内存,当可用内存不足时就会触发 mkl_malloc: failed to allocate memory

解决方案:

  1. 启用 CUDA 加速(推荐):如果您有 NVIDIA 显卡,请先安装 CUDA 12.8+ 和 cuDNN 9.11+,然后在主界面勾选“CUDA 加速”。GPU 可大幅降低内存需求并提升速度。
  2. 使用更小的模型:在无 GPU 或内存有限的情况下,将识别模型改为 mediumsmall(主界面下拉选择),large-v3-turbo 在 CPU 模式下内存占用极高。
  3. 释放系统内存:关闭其他占用内存的软件(如浏览器、大型应用),检查虚拟内存设置,确保有足够的页面文件空间。
  4. 调整高级选项:菜单栏 → 工具/选项 → 高级选项 → 语音识别参数:

    • CUDA数据类型 设为 int8(即使未用 CUDA,此设置也会影响 CPU 量化,降低内存占用)
    • 减少 beam_sizebest_of 为 1
  5. 重启电脑:清空内存缓存后再试。

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