Method forward

funasr/models/transducer/model.py:190–274

Encoder + Decoder + Calc loss Args: speech: (Batch, Length, ...) speech_lengths: (Batch, ) text: (Batch, Length) text_lengths: (Batch,)

(
        self,
        speech: torch.Tensor,
        speech_lengths: torch.Tensor,
        text: torch.Tensor,
        text_lengths: torch.Tensor,
        **kwargs,
    )

        self.ctc_weight = 0.0

    def forward(
        self,
        speech: torch.Tensor,
        speech_lengths: torch.Tensor,
        text: torch.Tensor,
        text_lengths: torch.Tensor,
        **kwargs,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Encoder + Decoder + Calc loss
        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
                text: (Batch, Length)
                text_lengths: (Batch,)
        """
        if len(text_lengths.size()) > 1:
            text_lengths = text_lengths[:, 0]
        if len(speech_lengths.size()) > 1:
            speech_lengths = speech_lengths[:, 0]

        batch_size = speech.shape[0]
        # 1. Encoder
        encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)
        if (
            hasattr(self.encoder, "overlap_chunk_cls")
            and self.encoder.overlap_chunk_cls is not None
        ):
            encoder_out, encoder_out_lens = self.encoder.overlap_chunk_cls.remove_chunk(
                encoder_out, encoder_out_lens, chunk_outs=None
            )
        # 2. Transducer-related I/O preparation
        decoder_in, target, t_len, u_len = get_transducer_task_io(
            text,
            encoder_out_lens,
            ignore_id=self.ignore_id,
        )

        # 3. Decoder
        self.decoder.set_device(encoder_out.device)
        decoder_out = self.decoder(decoder_in, u_len)

        # 4. Joint Network
        joint_out = self.joint_network(encoder_out.unsqueeze(2), decoder_out.unsqueeze(1))

        # 5. Losses
        loss_trans, cer_trans, wer_trans = self._calc_transducer_loss(
            encoder_out,
            joint_out,
            target,
            t_len,
            u_len,
        )

        loss_ctc, loss_lm = 0.0, 0.0

        if self.use_auxiliary_ctc:
            loss_ctc = self._calc_ctc_loss(
                encoder_out,