[BUG] 带有参考文本的 Transcribe 无法自动识别语言

[zhzLuke96]

现象

transcribe 的时候，如果 language 为 auto ，并且带有参考脚本的时候，报错

Transcribe:  22%|████████████                                            | 1.3/6.04 [00:01<00:04,  1.15sec/s]
Traceback (most recent call last):
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/queueing.py", line 536, in process_events
    response = await route_utils.call_process_api(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/route_utils.py", line 322, in call_process_api
    output = await app.get_blocks().process_api(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1935, in process_api
    result = await self.call_function(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1518, in call_function
    prediction = await fn(*processed_input)
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/utils.py", line 793, in async_wrapper
    response = await f(*args, **kwargs)
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/utils.py", line 793, in async_wrapper
    response = await f(*args, **kwargs)
  File "/[hidden]/modules/webui/asr_tabs/transcribe_tab.py", line 163, in submit
    audio = await handler.enqueue()
  File "/[hidden]/modules/core/handler/STTHandler.py", line 35, in enqueue
    result = self.chunker.transcribe(audio=self.input_audio, config=self.stt_config)
  File "/[hidden]/modules/core/models/stt/STTChunker.py", line 285, in transcribe
    result = self.transcribe_to_result(audio, config)
  File "/[hidden]/modules/core/models/stt/STTChunker.py", line 216, in transcribe_to_result
    result = self.model.transcribe_to_result(chunk.audio, config)
  File "/[hidden]/modules/core/models/stt/Whisper.py", line 249, in transcribe_to_result
    result = self.force_align(audio=audio, config=config)
  File "/[hidden]/modules/core/models/stt/Whisper.py", line 339, in force_align
    aligned_result = model.align(
  File "/[hidden]/venv/lib/python3.10/site-packages/stable_whisper/alignment.py", line 184, in align
    tokenizer, supported_languages = get_alignment_tokenizer(model, is_faster_model, text, language, tokenizer)
  File "/[hidden]/venv/lib/python3.10/site-packages/stable_whisper/alignment.py", line 377, in get_alignment_tokenizer
    raise TypeError('expected argument for language')
TypeError: expected argument for language
Traceback (most recent call last):
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/queueing.py", line 536, in process_events
    response = await route_utils.call_process_api(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/route_utils.py", line 322, in call_process_api
    output = await app.get_blocks().process_api(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1935, in process_api
    result = await self.call_function(
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1518, in call_function
    prediction = await fn(*processed_input)
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/utils.py", line 793, in async_wrapper
    response = await f(*args, **kwargs)
  File "/[hidden]/venv/lib/python3.10/site-packages/gradio/utils.py", line 793, in async_wrapper
    response = await f(*args, **kwargs)
  File "/[hidden]/modules/webui/asr_tabs/transcribe_tab.py", line 163, in submit
    audio = await handler.enqueue()
  File "/[hidden]/modules/core/handler/STTHandler.py", line 35, in enqueue
    result = self.chunker.transcribe(audio=self.input_audio, config=self.stt_config)
  File "/[hidden]/modules/core/models/stt/STTChunker.py", line 285, in transcribe
    result = self.transcribe_to_result(audio, config)
  File "/[hidden]/modules/core/models/stt/STTChunker.py", line 216, in transcribe_to_result
    result = self.model.transcribe_to_result(chunk.audio, config)
  File "/[hidden]/modules/core/models/stt/Whisper.py", line 249, in transcribe_to_result
    result = self.force_align(audio=audio, config=config)
  File "/[hidden]/modules/core/models/stt/Whisper.py", line 339, in force_align
    aligned_result = model.align(
  File "/[hidden]/venv/lib/python3.10/site-packages/stable_whisper/alignment.py", line 184, in align
    tokenizer, supported_languages = get_alignment_tokenizer(model, is_faster_model, text, language, tokenizer)
  File "/[hidden]/venv/lib/python3.10/site-packages/stable_whisper/alignment.py", line 377, in get_alignment_tokenizer
    raise TypeError('expected argument for language')

如果 language不为auto 就不会报错，应该是没有自动识别

预期

支持自动识别语言