#8784 Error during speech recognition stage[faster-whisper(Built-in)] unsupported device cuda:3,Traceback (most recent call la

103.159**5 Posted at: 3 hours ago

Error during speech recognition stage[faster-whisper(Built-in)] unsupported device cuda:3,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
ValueError: unsupported device cuda:3

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
ValueError: unsupported device cuda:3

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 67, 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
ValueError: unsupported device cuda:3

Traceback (most recent call last):
File "videotrans\task\only_one.py", line 84, 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\con
......
eo_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4,
self.cfg.cache_folder='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221'
Not-Selected Clear Output, Selected Enable CUDA
STT Service:faster-whisper(Built-in), large-v3, Spoken: English
Max Speech Duration (s): 5s
Min Speech Duration (ms): 3000ms
Min Silence Duration for Split (ms): 600ms
VAD Speech Threshold: 0.45
VAD Engine: tenvad
Compute Type:int8_bfloat16
Beam Size:5
Best Of:5
Condition on Previous Text:False
Repetition Penalty:1.0
Compression Ratio Threshold:2.4
TranslateSrt:Google Translate (Free),Spoken:English,Target:Burmese
Traditional Translation Batch Size (Lines): 5
TTS Service:Edge-TTS (Free), Voices:Thiha(Male/MM), Target:Burmese
Volume:+0%, Speed:+0%, Selected TTS Speedup, Not-Selected Slow Video
Selected Trim Start/End Silence in Dubbed Segments
Selected Strip All Silence in Dubbed Segments
Silence Removal Intensity: max
No Subtitles
Selected Output MP4 Only
video duration:848166ms
video information:{'video_fps': 30.0, 'r_frame_rate': 30.0, 'video_codec_name': 'h264', 'audio_codec_name': 'aac', 'width': 1920, 'height': 1080, 'time': 848166, 'streams_len': 2, 'streams_audio': 1, 'video_streams': 1, 'color': 'yuv420p'}
self.cfg=TaskCfgVTT(uuid='6084d96221', name='C:/Users/welcome/Downloads/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264).mp4', dirname='C:/Users/welcome/Downloads', noextname='Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)', basename='Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264).mp4', ext='mp4', target_dir='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4', cache_folder='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221', is_cuda=True, source_language='English', source_language_code='en', source_sub='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4/en.srt', source_wav='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221/en.wav', source_wav_output='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4/en.m4a', target_language='Burmese', target_language_code='my', target_sub='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4/my.srt', target_wav='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221/target.wav', target_wav_output='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4/my.m4a', translate_type=0, tts_type=0, volume='+0%', pitch='+0Hz', voice_rate='+0%', voice_role='Thiha(Male/MM)', voice_autorate=True, video_autorate=False, remove_silent_mid=False, align_sub_audio=True, detect_language='en', recogn_type=0, model_name='large-v3', shibie_audio=None, remove_noise=False, enable_diariz=False, nums_diariz=0, rephrase=False, fix_punc=0, app_mode='biaozhun', subtitle_language=None, targetdir_mp4='C:/Users/welcome/Downloads/_video_out/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264)-mp4/Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264).mp4', novoice_mp4='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221/novoice.mp4', is_separate=False, embed_bgm=True, instrument='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221/instrument.wav', vocal='C:/Users/welcome/Downloads/win-pyvideotrans-v4.18/tmp/10804/6084d96221/vocal.wav', clear_cache=False, background_music='', subtitle_type=0, only_out_mp4=True, recogn2pass=False, output_srt=0, copysrt_rawvideo=False, loop_backaudio=1, backaudio_volume=0.8, batch=False)
video_duration_ms=848166

=
system:Windows-10-10.0.26300-SP0
version:v4.18
frozen:True
language:en_US
root_dir:C:/Users/welcome/Downloads/win-pyvideotrans-v4.18

1 Replies

#1 DeepSeek reply at 3 hours ago

The core cause is that the software passed the GPU index as the device string cuda:3 to faster-whisper/ctranslate2. This build only accepts cuda plus a separate device_index, so it throws ValueError: unsupported device cuda:3. In other words, this is not a VRAM or model problem, but a device-index assignment problem (the machine has multiple GPUs, or the GPU enumeration assigned index 3, which this CUDA/ctranslate2 combination cannot use).

Solution (in order, stop once it works):

  1. Quickest workaround: uncheck CUDA Acceleration on the main window and run again with CPU. Everything else can stay as is.
  2. Force the device index back to 0: open Menu → Tools/Options → Advanced Options → General → Force Compute Device for Heavy Tasks, and set it to cuda:0 (or the index of your working NVIDIA card). Then re-run. Do not leave it at a non-existent or wrong index.
  3. Confirm the hardware environment: only NVIDIA GPUs support CUDA. Update the NVIDIA driver to the latest version, and make sure CUDA 12.8 + cuDNN 9.x are installed, with the cuDNN bin/lib directories in the system PATH. If CUDA is not properly installed, the device string will be rejected.
  4. Clean up the input file: the current file name is very long and contains —, parentheses, commas, and spaces (Stop Paying for AI Videos — 5 Free Unlimited AI Generators, No Watermark (Save Money) - AIWaliSarkar (1080p, h264).mp4). Rename it to a short name with only English letters and digits (e.g. video1.mp4) in a shallow directory such as D:/videos, then re-import.
  5. If it still reports unsupported device cuda:3 after steps 2–4, restart the software (to rebuild the process pool) and send the last ~30 lines of the error from the latest logs/YYYYMMDD.log file in the software directory for analysis.

Additional note: the translation channel is Google Translate (Free), which requires a proxy in mainland China; the TTS voice Thiha and the target language Burmese are fine.

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