funasr/export/models/encoder/conformer_encoder.py
@@ -61,7 +61,6 @@ speech: torch.Tensor, speech_lengths: torch.Tensor, ): speech = speech * self._output_size ** 0.5 mask = self.make_pad_mask(speech_lengths) mask = self.prepare_mask(mask) if self.embed is None:
