无法正常使用显卡cuda运行任务 #809
Description
问题描述 | Problem Description
我的是5060ti 16g显卡，驱动版本580.97，cuda环境版本是12.1，在转录时只能用cpu运行，无法用cuda运行，一开始报错：An error occured while synchronizing the model Systran/faster-whisper-tiny from the Hugging Face Hub:。后来直接报错无信息了。
另外还有个问题，如何使用关键词？每次输入也是报错。不知道是格式还是组件问题。我希望将我的专业词汇加入纠正过程。
日志信息（可选）| Logs (Optional)
===========转录任务开始===========
时间：2025-09-12 19:59:41.244958
开始转换音频
转换为音频执行命令: ffmpeg -i D:\FFOutput\111.m4a -map 0:a -ac 1 -ar 16000 -af aresample=async=1 -y C:\Users\Administrator\AppData\Local\Temp\tmpdw01nt1m.wav
开始语音转录
Faster Whisper 执行命令: faster-whisper-xxl -m large-v3 --print_progress --model_dir D:\AI\卡卡字幕助手\VideoCaptioner\AppData\models C:\Users\Administrator\AppData\Local\Temp\bk_asr\tmp5aicjk93\audio.wav -l zh -d cuda --output_format srt -o source --vad_filter true --vad_threshold 0.01 --vad_method silero_v4 --one_word 0 --sentence --max_line_width 30 --max_line_count 1 --max_comma 20 --max_comma_cent 50 --beep_off
Standalone Faster-Whisper-XXL r245.2 running on: CUDA
Note: 'large-v3' model may produce worse results than 'large-v2'!
Starting sequential inference to transcribe: C:\Users\Administrator\AppData\Local\Temp\bk_asr\tmp5aicjk93\audio.wav
Traceback (most recent call last):
File "D:\whisper-fast-XXL\__main__.py", line 2245, in <module>
File "D:\whisper-fast-XXL\__main__.py", line 2118, in cli
File "faster_whisper\transcribe.py", line 2179, in restore_speech_timestamps
File "faster_whisper\transcribe.py", line 1471, in generate_segments
File "faster_whisper\transcribe.py", line 1719, in encode
RuntimeError: cuBLAS failed with status CUBLAS_STATUS_NOT_SUPPORTED
[PYI-33960:ERROR] Failed to execute script '__main__' due to unhandled exception!
Faster Whisper 返回值: 1
2025-09-12 19:59:53 - faster_whisper - ERROR - Faster Whisper 错误:
2025-09-12 19:59:53 - transcript_thread - ERROR - 转录过程中发生错误:
Traceback (most recent call last):
File "D:\AI\卡卡字幕助手\VideoCaptioner\app\thread\transcript_thread.py", line 101, in run
asr_data = transcribe(
^^^^^^^^^^^
File "D:\AI\卡卡字幕助手\VideoCaptioner\app\core\bk_asr\transcribe.py", line 86, in transcribe
asr_data = asr.run(callback=callback)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\卡卡字幕助手\VideoCaptioner\app\core\bk_asr\base.py", line 54, in run
resp_data = self._run(callback, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\卡卡字幕助手\VideoCaptioner\app\core\bk_asr\faster_whisper.py", line 266, in _run
raise RuntimeError(error_msg)
RuntimeError