#7132 语音识别阶段出错[faster-whisper(内置)] Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64:

103.144**0 Posted at: 1 day ago

语音识别阶段出错[faster-whisper(内置)] Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64:Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 164, in faster_whisper
File "faster_whisper\transcribe.py", line 916, in transcribe
File "faster_whisper\feature_extractor.py", line 215, in call
File "faster_whisper\feature_extractor.py", line 189, in stft
File "numpy\fft\_pocketfft.py", line 409, in rfft
File "numpy\fft\_pocketfft.py", line 70, in _raw_fft
numpy.core._exceptions._ArrayMemoryError: Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64

Traceback (most recent call last):

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

File "videotrans\task\job.py", line 102, in process_task

File "videotrans\task\speech2text.py", line 100, in recogn

File "videotrans\recognition\__init__.py", line 211, in run

File "videotrans\recognition\_base.py", line 98, in run

File "videotrans\recognition\_whisper.py", line 41, in _exec

File "videotrans\recognition\_whisper.py", line 130, in _faster

File "videotrans\configure\base.py", line 270, in _new_process

videotrans.configure.excepts.VideoTransError: Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64:Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 164, in faster_whisper
File "faster_whisper\transcribe.py", line 916, in transcribe
File "faster_whisper\feature_extractor.py", line 215, in call
File "faster_whisper\feature_extractor.py", line 189, in stft
File "numpy\fft\_pocketfft.py", line 409, in rfft
File "numpy\fft\_pocketfft.py", line 70, in _raw_fft
numpy.core._exceptions._ArrayMemoryError: Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64
[SpeechToText]语音转录: self.out_format='srt',self.copysrt_rawvideo=False,self.spk_insert=True
[TaskCfgSTT]
原始输入文件名: D:/Study_Materials/课程素材/音频/接地1.mp3,
输出结果保存到文件夹: D:/Software/pyvideotrans/output/recogn,
临时文件夹: D:/Software/pyvideotrans/tmp/28008/328fe92697
已启用CUDA加速
未选降噪
语音识别:faster-whisper(内置), large-v3-turbo 断句方式:默认断句
发音语言:ru
未选降噪
代理地址:http://127.0.0.1:7897

=
system:Windows-10-10.0.22631-SP0
version:v4.10
frozen:True
language:zh
root_dir:D:/Software/pyvideotrans

#1 DeepSeek reply at 1 day ago

这是显存或内存不足导致的记忆体分配失败，常见于使用了 large-v3-turbo 大模型且显卡显存不够，或计算数据类型占用过高。

请依次尝试以下操作：

降低模型尺寸：在语音识别渠道中将模型从 large-v3-turbo 改为 medium 或 small。large-v3 系列基础运行需要约 6-8GB 显存。
调整计算数据类型：进入 菜单 → 工具 → 高级选项 → 语音识别参数，将 计算数据类型 由默认改为 int8（最省显存）或 float16（适合显卡）。
关闭上下文感知：在同样的高级选项中将 启用上下文感知 设为不勾选。
降低识别精度参数：将 beam_size 和 best_of 都调为 1。
若显存仍不足：可在确认 CUDA 配置正确的前提下，暂时关闭 启用CUDA加速，改用 CPU 运行（速度会慢但可正常处理）。
清理临时残留：关闭软件后，手动清空软件目录下 tmp/ 文件夹内的所有内容，避免残留缓存占用内存。

如果以上调整后仍然报错，请检查显卡是否可用的显存确实足够，并关闭其他占用显存的程序。

#7132 语音识别阶段出错[faster-whisper(内置)] Unable to allocate 1.33 GiB for an array with shape (1, 444787, 400) and data type float64:

1 Replies

请查阅相关文档：

Post Your Reply

Related resource