#8505 Unable to allocate 463. MiB for an array with shape (1, 151635, 400) and data type float64:Traceback (most recent call l

151.46**4 Posted at: 2 hours ago

Unable to allocate 463. MiB for an array with shape (1, 151635, 400) and data type float64:Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 121, 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 463. MiB for an array with shape (1, 151635, 400) and data type float64

Traceback (most recent call last):
File "videotrans\task\only_one.py", line 71, in run
File "videotrans\task\_stage_recogn.py", line 64, in recogn
File "videotrans\recognition\__init__.py", line 148, in run
File "videotrans\recognition\_base.py", line 101, in run
File "videotrans\recognition\_whisper.py", line 39, in _exec
File "videotrans\recognition\_whisper.py", line 136, in _faster
File "videotrans\configure\base.py", line 277, in _new_process
videotrans.configure.excepts.VideoTransError: Unable to allocate 463. MiB for an array with shape (1, 151635, 400) and data type float64:Traceback (most recent call last):
File "videotrans\process\stt_faster.py", line 121, 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 463. MiB for an array with shape (1, 151635, 400) and data type float64

trk=[TransCreate]: uuid='ebf1735851', proxy_str=None, cfg=[TaskCfgVTT]当前工作模式: 翻译视频 单视频模式
原始输入文件名: C:/Users/navar/Downloads/fundamentals-2.mp4,
输出结果保存到文件夹: C:/Users/navar/Downloads/_video_out/fundamentals-2-mp4,
临时文件夹: D:/pyVideoTrans/tmp/13564/ebf1735851
已选 清理已存在
未启用CUDA加速
STT Service:faster-whisper(Built-in), large-v3-turbo, 发音语言: English, , 数据类型:default
TranslateSrt:Google Translate (Free),原始语言:English,目标语言:Italian, 已选 SendSRT, settings.get('trans_thread')=5, settings.get('aitrans_thread')=40
TTS Service:Edge-TTS (Free), 角色:GiuseppeMultilingual(Male/IT), 配音语言:Italian
音量:+0%, 语速:+0%, 已选 音频加速, 未选 视频慢速
已选 移除每条字幕配音开头和结尾静音缓冲
静音移除力度: default
字幕: Hard Subtitles , precent=4, hasend=False, should_recogn=True, should_trans=True, should_dubbing=True, should_hebing=True, is_copy_video=True, video_codec_num=264, ignore_align=False, is_audio_trans=False, clone_ref='', should_recogn2=False, do_diarize=True
video_duration_ms=1899166

=
system:Windows-10-10.0.19045-SP0
version:v4.17
frozen:True
language:en_US
root_dir:D:/pyVideoTrans

1 Replies

#1 DeepSeek reply at 2 hours ago

This is not a CUDA or VRAM error: it is a system RAM (physical memory) shortage error. numpy needed to allocate a contiguous ~463 MiB array for the STFT spectrogram and failed. The array shape (1, 151635, 400) corresponds to roughly 25 minutes of 16 kHz audio being converted to a spectrogram in one pass, so the memory demand grows with the length of the audio (your file is 31:39 long) and was larger than the free RAM available at that moment.

Solution (in this order):

  1. Free up memory before starting: close browsers, other large programs and any second instance of pyVideoTrans; open Task Manager and confirm there is several GB of free memory before clicking Start Execution.

  2. Open Menu → Tools/Options → Advanced Options → General:

    • Max Concurrent CPU Tasks [Restart Required] = 1
    • Max Concurrent GPU Tasks [Restart Required] = 1
      Then restart the software for these to take effect (this caps parallel tasks and therefore peak memory).
  3. Split the long video into shorter parts (e.g. 5–10 minutes each) and translate them separately; the STFT memory requirement scales with audio length, so shorter inputs avoid the allocation entirely.
  4. Optional, further reduce memory pressure (this does not shrink the spectrogram itself but lowers overall usage): in the STT row, change the model from large-v3-turbo to medium or small, and clear Separate Vocal/BGM if it is checked.
  5. Make sure Clear Output is checked before retrying the same file, so stale intermediate files from the failed run are not reused.

Note the log shows 未启用CUDA加速 (CUDA not enabled), so GPU acceleration is not a factor here; enabling it would not fix a CPU RAM allocation failure.

If the error still occurs after the above, please provide the last ~30 lines of the latest log file in the logs/ folder under the software installation directory (D:/pyVideoTrans/logs/) so the actual memory state at the failure point can be analyzed.

Please refer to the related documentation:

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