funasr/bin/vad_infer.py
@@ -175,7 +175,8 @@ batch_size = speech.shape[0] segments = [[]] * batch_size if self.frontend is not None: feats, feats_len = self.frontend.forward(speech, speech_lengths, is_final) reset = in_cache == dict() feats, feats_len = self.frontend.forward(speech, speech_lengths, is_final, reset) fbanks, _ = self.frontend.get_fbank() else: raise Exception("Need to extract feats first, please configure frontend configuration")
