fix bug for contextual train

shixian.shi

2023-11-24 34b2682fba92bf4f8e163d22094e7de57e3290ca

fix bug for contextual train

 funasr/models/e2e_asr_contextual_paraformer.py

@@ -207,7 +207,7 @@
        # 2b. Attention decoder branch
        if self.ctc_weight != 1.0:
            loss_att, acc_att, cer_att, wer_att, loss_pre, loss_ideal = self._calc_att_clas_loss(
                encoder_out, encoder_out_lens, text, text_lengths, hotword_pad, hotword_lengths, ideal_attn
                encoder_out, encoder_out_lens, text, text_lengths, hotword_pad, hotword_lengths
            )

        # 3. CTC-Att loss definition
@@ -242,7 +242,6 @@
            ys_pad_lens: torch.Tensor,
            hotword_pad: torch.Tensor,
            hotword_lengths: torch.Tensor,
            ideal_attn: torch.Tensor,
    ):
        encoder_out_mask = (~make_pad_mask(encoder_out_lens, maxlen=encoder_out.size(1))[:, None, :]).to(
            encoder_out.device)