#8507 语音识别阶段出错[faster-whisper(内置)] mkl_malloc: failed to allocate memory,Traceback (most recent call last): File

108.181**1 Posted at: 1 hour ago

语音识别阶段出错[faster-whisper(内置)] mkl_malloc: failed to allocate memory,Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 84, in faster_whisper
File "videotrans\process\stt_faster.py", line 73, in _create_model
File "videotrans\process\stt_faster.py", line 77, in _create_model
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\trans
......
last):
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 84, in faster_whisper
File "videotrans\process\stt_faster.py", line 73, in _create_model
File "videotrans\process\stt_faster.py", line 77, in _create_model
File "videotrans\process\stt_faster.py", line 54, in _create_model
File "faster_whisper\transcribe.py", line 689, in init
RuntimeError: mkl_malloc: failed to allocate memory

self.cfg=[TaskCfgVTT]当前工作模式: 转录并翻译字幕
原始输入文件名: D:/Downloads/Videos/!Submaker/exploitedcollegegirls.26.10.15.carly.kiss.and.honey.april.i.got.to.lick.it.out.of.her.xxx.mp4,
输出结果保存到文件夹: D:/Downloads/Videos/!Submaker/exploitedcollegegirls.26.10.15.carly.kiss.and.honey.april.i.got.to.lick.it.out.of.her.xxx-mp4,
临时文件夹: D:/program/videotrans/tmp/26612/ab9faecda6
已选 清理已存在
未启用CUDA加速
未选 降噪
语音识别:faster-whisper(内置), large-v3, 发音语言: 英语, 断句方式:LLM重新断句
翻译渠道:DeepSeek,原始语言:英语,目标语言:简体中文, 未选 发送完整字幕
输出字幕格式: 单字幕
代理地址:http://127.0.0.1:7897


=
system:Windows-10-10.0.26300-SP0
version:v4.14
frozen:True
language:zh_CN
root_dir:D:/program/videotrans

1 Replies

#1 DeepSeek reply at 1 hour ago

核心原因是内存不足:在未启用 CUDA 加速时,faster-whisper 使用 large-v3 模型运行需要消耗大量内存,当前系统可用内存不足以完成模型加载,因此触发 mkl_malloc: failed to allocate memory。

按以下顺序处理:

  1. 降低模型体积(最有效):主界面将语音识别模型从 large-v3 改为 medium、small 或 base,large-v3 在纯 CPU 模式下的内存占用远超普通机型承载能力。
  2. 更改计算数据类型:菜单 → 工具/选项 → 高级选项 → 语音识别参数 → 计算数据类型 改为 int8(消耗资源最少、速度最快)。
  3. 减小搜索深度:同区域将 识别准确度beam_size 与 best_of 从默认 5 调低为 1。
  4. 关闭上下文感知:同区域将 启用上下文感知 设为否,减少额外内存占用。
  5. 限制并发:高级选项 → 通用设置 → CPU同时任务数 设为 1,批量翻译视频时每批数量 设为 1;GPU同时任务数 设为 1。
  6. 释放系统内存:关闭浏览器等其他占内存程序;确认系统虚拟内存(页面文件)未被禁用或设置过小。
  7. 如有 NVIDIA 显卡:安装 CUDA 12.8 + cuDNN 9,主界面勾选 CUDA 加速 后可显著缓解 CPU 内存压力(前提是显存足够跑所选模型)。

提示:你当前选择的是「LLM重新断句」,该功能本身走在线 API 不占本地内存,问题仍在模型加载环节,按上述 1~6 步处理后重新勾选 清理已生成 再执行。

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