File "/usr/local/lib/python3.11/dist-packages/transformers/pipelines/pt_utils.py", line 271, in next processed = self.infer(next(self.iterator), **self.params) ^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/torch/utils/data/dataloader.py", line 741, in next data = self._next_data() ^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/torch/utils/data/dataloader.py", line 801, in _next_data data = self._dataset_fetcher.fetch(index) # may raise StopIteration ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/torch/utils/data/_utils/fetch.py", line 35, in fetch data.append(next(self.dataset_iter)) ^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/transformers/pipelines/pt_utils.py", line 188, in next processed = next(self.subiterator) ^^^^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/transformers/pipelines/automatic_speech_recognition.py", line 486, in preprocess processed = self.feature_extractor( ^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.11/dist-packages/transformers/models/lasr/feature_extraction_lasr.py", line 266, in call input_features = self._torch_extract_fbank_features(input_features, device, center) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ TypeError: LasrFeatureExtractor._torch_extract_fbank_features() takes from 2 to 3 positional arguments but 4 were given

Fix Plan

Update transformers version

The issue is introduced in transformers version 5.2.0. We can try downgrading to a previous version that does not have this issue.

Update code to match new transformers version

If downgrading is not an option, we need to update the code to match the new version of transformers. The error message indicates that the LasrFeatureExtractor._torch_extract_fbank_features() method now takes 2 to 3 positional arguments, but 4 were given.

Here's an example of how to update the code:

import torch
from transformers import pipeline

# Create a pipeline for ASR
pipe = pipeline("automatic-speech-recognition", model=model_id)

# Update the audio processing code to match the new transformers version
def process_audio(audio):
    # Assuming audio is a tensor
    if len(audio.shape) == 1:  # mono audio
        audio = torch.unsqueeze(audio, 0)  # add batch dimension
    elif len(audio.shape) == 2:  # stereo audio
        pass  # no need to add batch dimension
    else:
        raise ValueError("Invalid audio shape")

    # Process audio using the pipeline
    result = pipe(audio, chunk_length_s=20, stride_length_s=2)
    return result

Verify the fix

To verify that the fix worked, you can run the updated code and check if the error is resolved. You can also add some print statements to verify that the audio is being processed correctly.

print("Audio shape:", audio.shape)
result = process_audio(audio)
print("Result:", result)

Extra Tips

• Make sure to check the transformers documentation for any changes in the API that may affect your code.
• If you're using a custom dataset or task, make sure to update your code to match the new transformers version.
• Consider using a virtual environment to