#8586 load_backend: loaded CPU backend from C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\whisper-cpp\ggml-cpu-haswell.

27.116**6 Posted at: 2 hours ago

load_backend: loaded CPU backend from C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\whisper-cpp\ggml-cpu-haswell.dll
error: failed to initialize whisper context

Traceback (most recent call last):
File "videotrans\recognition\_cpp.py", line 121, in _exec
videotrans.configure.excepts.SpeechToTextError: load_backend: loaded CPU backend from C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\whisper-cpp\ggml-cpu-haswell.dll
error: failed to initialize whisper context

During handling of the above exception, another exception occurred:

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\_cpp.py", line 128, in _exec
videotrans.configure.excepts.SpeechToTextError: load_backend: loaded CPU backend from C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\whisper-cpp\ggml-cpu-haswell.dll
error: failed to initialize whisper context

trk=[TransCreate]: uuid='d82d98f4c3', 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/seenu/Downloads/@绘春枝 s personal homepage.mp4,
输出结果保存到文件夹: C:/Users/seenu/Downloads/_video_out/@绘春枝 s personal homepage-mp4,
临时文件夹: C:/Users/seenu/Downloads/win-pyvideotrans-v4.17-1028/tmp/25788/d82d98f4c3
已选清理已存在
未启用CUDA加速
STT Service:Whisper.cpp(WinBuilt-in), , 发音语言: Simplified Chinese, ,
TranslateSrt:Microsoft Translate (Free),原始语言:Simplified Chinese,目标语言:English, 已选 SendSRT, settings.get('trans_thread')=1, settings.get('aitrans_thread')=40
未选配音角色，不进行配音
字幕: Hard Subtitles
已选仅输出mp4, precent=4, hasend=False, should_recogn=True, should_trans=True, should_dubbing=False, 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=148366

=
system:Windows-10-10.0.26300-SP0
version:v4.17
frozen:True
language:en_US
root_dir:C:/Users/seenu/Downloads/win-pyvideotrans-v4.17-1028

#1 DeepSeek reply at 2 hours ago

The CPU backend loaded fine — the failure is that whisper.cpp could not load its model file (failed to initialize whisper context is a model-load failure, not a missing DLL). Typical causes: the .bin ggml model was never downloaded, was downloaded incompletely/corrupted, or a non-ggml model file (e.g. a .pt/CT2 model) was placed there.

Fix, in order:

Open C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\models\ and confirm the ggml model file for the model selected on the main UI exists (e.g. ggml-large-v3-turbo.bin / ggml-medium.bin), and that its size is roughly the expected size (hundreds of MB to ~3 GB). If it is missing, 0 KB, or much smaller than expected, delete it and re-download.
Download the correct file from https://huggingface.co/ggerganov/whisper.cpp/tree/main (mirror: https://hf-mirror.com/ggerganov/whisper.cpp/tree/main) and put the .bin directly into the models folder. Do not use .pt (openai-whisper) or CT2 files here.
Rename the input video to a short name with only English letters/digits, e.g. video1.mp4 — the current name contains Chinese characters and @, which can also break command parsing in later stages.
Move the whole program out of C:\Users\...\Downloads\ into a short ASCII path such as D:\pyVideoTrans (Downloads/China paths and deep folders are a frequent source of failures).
If it still fails, switch STT Service to faster-whisper(Built-in) with model large-v3-turbo to bypass whisper.cpp entirely.

If the same error persists after the model file is verified complete, provide the last 30 lines of the latest log file in the logs/ folder of the installation directory.

#8586 load_backend: loaded CPU backend from C:\Users\seenu\Downloads\win-pyvideotrans-v4.17-1028\whisper-cpp\ggml-cpu-haswell.

1 Replies

Please refer to the related documentation:

Post Your Reply

Related resource