#3826 TaskCfgVTT(is_cuda=True, uuid='390e4e6473', cache_folder='G:/VideoTrans/Trans/win-pyvideotrans-v3.98-327/tmp/36364/390e4

103.172* Posted at: 3 hours ago 👁4

语音识别阶段出错 [faster-whisper(本地)] Traceback (most recent call last):
File "videotrans\process\stt_fun.py", line 346, in faster_whisper
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: CUDA failed with error out of memory

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 272, 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 105, 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 346, in faster_whisper
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: CUDA failed with error out of memory
TaskCfgVTT(is_cuda=True, uuid='390e4e6473', cache_folder='G:/VideoTrans/Trans/win-pyvideotrans-v3.98-327/tmp/36364/390e4e6473', target_dir='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4', source_language='泰国语', source_language_code='th', source_sub='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4/th.srt', source_wav='G:/VideoTrans/Trans/win-pyvideotrans-v3.98-327/tmp/36364/390e4e6473/th.wav', source_wav_output='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4/th.m4a', target_language='简体中文', target_language_code='zh-cn', target_sub='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4/zh-cn.srt', target_wav='G:/VideoTrans/Trans/win-pyvideotrans-v3.98-327/tmp/36364/390e4e6473/target.wav', target_wav_output='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4/zh-cn.m4a', name='C:/Users/lsl/Downloads/Namnung and Sunny tiktok live #namnung #sunny.mp4', noextname='Namnung and Sunny tiktok live #namnung #sunny', basename='Namnung and Sunny tiktok live #namnung #sunny.mp4', ext='mp4', dirname='C:/Users/lsl/Downloads', shound_del_name=None, translate_type=1, 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='th', recogn_type=0, model_name='large-v3', 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='C:/Users/lsl/Downloads/_video_out/Namnung and Sunny tiktok live #namnung #sunny-mp4/Namnung and Sunny tiktok live #namnung #sunny.mp4', novoice_mp4='G:/VideoTrans/Trans/win-pyvideotrans-v3.98-327/tmp/36364/390e4e6473/novoice.mp4', is_separate=False, instrument=None, vocal=None, back_audio='', clear_cache=True, background_music=None, subtitle_type=0, only_out_mp4=False, recogn2pass=True, output_srt=0, copysrt_rawvideo=False)

1 Replies

这是由于显卡显存不足导致的。错误信息 CUDA failed with error out of memory 表明在使用 faster-whisper(本地)large-v3 模型进行语音识别时,GPU 显存耗尽。

解决方案:

  1. 更换更小的语音识别模型:在主界面的语音识别渠道右侧,将模型从 large-v3 更换为 mediumsmallbaselarge-v3 模型至少需要 10GB 显存。
  2. 调整高级设置以降低显存占用:点击 菜单 -> 工具 -> 高级选项,在“语音识别参数”区域进行以下修改:

    • CUDA数据类型:将 float32 改为 float16int8
    • 识别准确度beam_size:将 5 改为 1
    • 识别准确度best_of:将 5 改为 1
    • 启用上下文感知:确保此项为 false(未勾选)。
  3. 关闭二次识别:在主界面右上角,取消勾选“二次识别”选项。同时翻译多个视频时,二次识别会显著增加显存需求。
  4. 检查多显卡配置:如果您有多块显卡,且第一块显卡显存较小,可以尝试升级到 v3.98-317 或更高版本,软件会自动选择可用显存最大的显卡。

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