Remove T.VAD from transform batch consistency test

[mthrok]

See #1348 for the background

T.VAD currently treats all the tracks in the input batch as belong to the same audio data. This means that it returns different results when samples are batched and samples are fed independently.

There is a test to ensure the consistency of operation result regardless of batched/un-bached, but T.VAD in this test does not make sense. We would like to remove it.

audio/test/torchaudio_unittest/transforms/torchscript_consistency_impl.py

Lines 92 to 95 in 6a677ac

	def test_Vad(self):
	filepath = common_utils.get_asset_path("vad-go-mono-32000.wav")
	waveform, sample_rate = common_utils.load_wav(filepath)
	self._assert_consistency(T.Vad(sample_rate=sample_rate), waveform)

Steps

1. Remove the test

Testing and building

Please refer to CONTRIBUTING.md.

[jcaw]

I'll take this, but doesn't the test you quoted test JIT consistency, rather than batch consistency? (Edit: so while T.VAD remains in the codebase, it would be desirable to keep.)

[mthrok]

Oops my bad. Yes, I meant the batch consistency test.