funasr/models/e2e_asr_paraformer.py
@@ -712,9 +712,9 @@ def calc_predictor_chunk(self, encoder_out, cache=None): pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index = \ pre_acoustic_embeds, pre_token_length = \ self.predictor.forward_chunk(encoder_out, cache["encoder"]) return pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index return pre_acoustic_embeds, pre_token_length def cal_decoder_with_predictor_chunk(self, encoder_out, sematic_embeds, cache=None): decoder_outs = self.decoder.forward_chunk(
