add docstring for functional_impl test

test/torchaudio_unittest/functional/functional_impl.py

@@ -584,6 +584,13 @@ def test_rnnt_loss_costs_and_gradients_random_data_with_numpy_fp32(self):
             self._test_costs_and_gradients(data=data, ref_costs=ref_costs, ref_gradients=ref_gradients)
 
     def test_mvdr_weights_rtf(self):
+        """Verify ``F.mvdr_weights_rtf`` method by numpy implementation.
+        Given the relative transfer function (RTF) of target speech (Tensor of dimension `(..., freq, channel)`),
+        the PSD matrix of noise (Tensor of dimension `(..., freq, channel, channel)`), and an integer
+        indicating the reference channel as inputs, ``F.mvdr_weights_rtf`` outputs the mvdr weights
+        (Tensor of dimension `(..., freq, channel)`), which should be close to the output of
+        ``mvdr_weights_rtf_numpy``.
+        """
         n_fft_bin = 10
         channel = 4
         reference_channel = 0
@@ -603,6 +610,13 @@ def test_mvdr_weights_rtf(self):
         )
 
     def test_mvdr_weights_rtf_with_tensor(self):
+        """Verify ``F.mvdr_weights_rtf`` method by numpy implementation.
+        Given the relative transfer function (RTF) of target speech (Tensor of dimension `(..., freq, channel)`),
+        the PSD matrix of noise (Tensor of dimension `(..., freq, channel, channel)`), and a one-hot Tensor
+        indicating the reference channel as inputs, ``F.mvdr_weights_rtf`` outputs the mvdr weights
+        (Tensor of dimension `(..., freq, channel)`), which should be close to the output of
+        ``mvdr_weights_rtf_numpy``.
+        """
         n_fft_bin = 10
         channel = 4
         reference_channel = np.zeros(channel)

torchaudio/functional/functional.py

@@ -1727,7 +1727,7 @@ def mvdr_weights_rtf(
             reference_channel = reference_channel.to(psd_n.dtype)
             scale = torch.einsum("...c,...c->...", [rtf.conj(), reference_channel[..., None, :]])
         else:
-            raise TypeError(f"Unsupported dtype for reference_channel. Found: {type(reference_channel)}.")
+            raise TypeError(f"Expected 'int' or 'Tensor' for reference_channel. Found: {type(reference_channel)}.")
 
         beamform_weights = beamform_weights * scale[..., None]