How do I use a bytes wav object instead of passing a referential string for a path to a wav file?

[Baquara]

Hello, consider this code:

def transcribe(file):
    torch.hub.download_url_to_file('https://opus-codec.org/static/examples/samples/speech_orig.wav',
                                dst ='speech_orig.wav', progress=True)
    test_files = glob('speech_orig.wav')
    batches = split_into_batches(test_files, batch_size=10)
    input = prepare_model_input(read_batch(batches[0]),
                                device=device)
    output = model(input)
    return decoder(output[0])

I want to use "file" (a bytes object) instead of "speech_orig.wav", how do I do it?

If I replace it in any of the steps I get errors.

[snakers4]

You can just override this function -

silero-models/src/silero/utils.py

Lines 22 to 37 in 20ce121

	def read_audio(path: str,
	target_sr: int = 16000):
	wav, sr = torchaudio.load(path)
	if wav.size(0) > 1:
	wav = wav.mean(dim=0, keepdim=True)
	if sr != target_sr:
	transform = torchaudio.transforms.Resample(orig_freq=sr,
	new_freq=target_sr)
	wav = transform(wav)
	sr = target_sr
	assert sr == target_sr
	return wav.squeeze(0)

[Baquara]

You can just override this function -

silero-models/src/silero/utils.py

Lines 22 to 37 in 20ce121

	def read_audio(path: str,
	target_sr: int = 16000):
	wav, sr = torchaudio.load(path)
	if wav.size(0) > 1:
	wav = wav.mean(dim=0, keepdim=True)
	if sr != target_sr:
	transform = torchaudio.transforms.Resample(orig_freq=sr,
	new_freq=target_sr)
	wav = transform(wav)
	sr = target_sr
	assert sr == target_sr
	return wav.squeeze(0)

I am sorry about asking this, but can you provide an example of how it can be done? Do I need to edit that file directly?