add paraformer online infer and finetune

haoneng.lhn

2023-05-26 84b4a01979ecc483096cccf5185dbe5e56946217

add paraformer online infer and finetune

 funasr/bin/asr_inference_launch.py

@@ -1618,6 +1618,8 @@
        return inference_uniasr(**kwargs)
    elif mode == "paraformer":
        return inference_paraformer(**kwargs)
    elif mode == "paraformer_online":
        return inference_paraformer(**kwargs)
    elif mode == "paraformer_streaming":
        return inference_paraformer_online(**kwargs)
    elif mode.startswith("paraformer_vad"):
@@ -1900,4 +1902,4 @@


if __name__ == "__main__":
    main()
    main()

 funasr/models/decoder/sanm_decoder.py

@@ -956,13 +956,13 @@
        """
        tgt = ys_in_pad
        tgt_mask = myutils.sequence_mask(ys_in_lens, device=tgt.device)[:, :, None]
        
        memory = hs_pad
        memory_mask = myutils.sequence_mask(hlens, device=memory.device)[:, None, :]
        if chunk_mask is not None:
            memory_mask = memory_mask * chunk_mask
            if tgt_mask.size(1) != memory_mask.size(1):
                memory_mask = torch.cat((memory_mask, memory_mask[:, -2:-1, :]), dim=1)

        memory = hs_pad
        memory_mask = myutils.sequence_mask(hlens, device=memory.device)[:, None, :]

        x = tgt
        x, tgt_mask, memory, memory_mask, _ = self.decoders(

 funasr/models/e2e_asr_paraformer.py

@@ -279,7 +279,7 @@

    def encode(
            self, speech: torch.Tensor, speech_lengths: torch.Tensor, ind: int = 0,
    ) -> Tuple[Tuple[Any, Optional[Any]], Any]:
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Frontend + Encoder. Note that this method is used by asr_inference.py
        Args:
                speech: (Batch, Length, ...)
@@ -649,7 +649,35 @@
        assert 0.0 <= ctc_weight <= 1.0, ctc_weight
        assert 0.0 <= interctc_weight < 1.0, interctc_weight

        super().__init__()
        super().__init__(
            vocab_size=vocab_size,
            token_list=token_list,
            frontend=frontend,
            specaug=specaug,
            normalize=normalize,
            preencoder=preencoder,
            encoder=encoder,
            postencoder=postencoder,
            decoder=decoder,
            ctc=ctc,
            ctc_weight=ctc_weight,
            interctc_weight=interctc_weight,
            ignore_id=ignore_id,
            blank_id=blank_id,
            sos=sos,
            eos=eos,
            lsm_weight=lsm_weight,
            length_normalized_loss=length_normalized_loss,
            report_cer=report_cer,
            report_wer=report_wer,
            sym_space=sym_space,
            sym_blank=sym_blank,
            extract_feats_in_collect_stats=extract_feats_in_collect_stats,
            predictor=predictor,
            predictor_weight=predictor_weight,
            predictor_bias=predictor_bias,
            sampling_ratio=sampling_ratio,
        )
        # note that eos is the same as sos (equivalent ID)
        self.blank_id = blank_id
        self.sos = vocab_size - 1 if sos is None else sos
@@ -705,6 +733,7 @@
        self.sampling_ratio = sampling_ratio
        self.criterion_pre = mae_loss(normalize_length=length_normalized_loss)
        self.step_cur = 0
        self.scama_mask = None
        if hasattr(self.encoder, "overlap_chunk_cls") and self.encoder.overlap_chunk_cls is not None:
            from funasr.modules.streaming_utils.chunk_utilis import build_scama_mask_for_cross_attention_decoder
            self.build_scama_mask_for_cross_attention_decoder_fn = build_scama_mask_for_cross_attention_decoder
@@ -859,7 +888,7 @@
        # Pre-encoder, e.g. used for raw input data
        if self.preencoder is not None:
            feats, feats_lengths = self.preencoder(feats, feats_lengths)

        
        # 4. Forward encoder
        # feats: (Batch, Length, Dim)
        # -> encoder_out: (Batch, Length2, Dim2)
@@ -1111,12 +1140,73 @@

        return sematic_embeds * tgt_mask, decoder_out * tgt_mask, pre_loss_att

    def calc_predictor(self, encoder_out, encoder_out_lens):

        encoder_out_mask = (~make_pad_mask(encoder_out_lens, maxlen=encoder_out.size(1))[:, None, :]).to(
            encoder_out.device)
        mask_chunk_predictor = None
        if self.encoder.overlap_chunk_cls is not None:
            mask_chunk_predictor = self.encoder.overlap_chunk_cls.get_mask_chunk_predictor(None,
                                                                                           device=encoder_out.device,
                                                                                           batch_size=encoder_out.size(
                                                                                               0))
            mask_shfit_chunk = self.encoder.overlap_chunk_cls.get_mask_shfit_chunk(None, device=encoder_out.device,
                                                                                   batch_size=encoder_out.size(0))
            encoder_out = encoder_out * mask_shfit_chunk
        pre_acoustic_embeds, pre_token_length, pre_alphas, pre_peak_index = self.predictor(encoder_out,
                                                                                           None,
                                                                                           encoder_out_mask,
                                                                                           ignore_id=self.ignore_id,
                                                                                           mask_chunk_predictor=mask_chunk_predictor,
                                                                                           target_label_length=None,
                                                                                           )
        predictor_alignments, predictor_alignments_len = self.predictor.gen_frame_alignments(pre_alphas[:, :-1],
                                                                                             encoder_out_lens)

        scama_mask = None
        if self.encoder.overlap_chunk_cls is not None and self.decoder_attention_chunk_type == 'chunk':
            encoder_chunk_size = self.encoder.overlap_chunk_cls.chunk_size_pad_shift_cur
            attention_chunk_center_bias = 0
            attention_chunk_size = encoder_chunk_size
            decoder_att_look_back_factor = self.encoder.overlap_chunk_cls.decoder_att_look_back_factor_cur
            mask_shift_att_chunk_decoder = self.encoder.overlap_chunk_cls.\
                get_mask_shift_att_chunk_decoder(None,
                                                 device=encoder_out.device,
                                                 batch_size=encoder_out.size(0)
                                                 )
            scama_mask = self.build_scama_mask_for_cross_attention_decoder_fn(
                predictor_alignments=predictor_alignments,
                encoder_sequence_length=encoder_out_lens,
                chunk_size=1,
                encoder_chunk_size=encoder_chunk_size,
                attention_chunk_center_bias=attention_chunk_center_bias,
                attention_chunk_size=attention_chunk_size,
                attention_chunk_type=self.decoder_attention_chunk_type,
                step=None,
                predictor_mask_chunk_hopping=mask_chunk_predictor,
                decoder_att_look_back_factor=decoder_att_look_back_factor,
                mask_shift_att_chunk_decoder=mask_shift_att_chunk_decoder,
                target_length=None,
                is_training=self.training,
            )
        self.scama_mask = scama_mask

        return pre_acoustic_embeds, pre_token_length, pre_alphas, pre_peak_index

    def calc_predictor_chunk(self, encoder_out, cache=None):

        pre_acoustic_embeds, pre_token_length = \
            self.predictor.forward_chunk(encoder_out, cache["encoder"])
        return pre_acoustic_embeds, pre_token_length

    def cal_decoder_with_predictor(self, encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens):
        decoder_outs = self.decoder(
            encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens, self.scama_mask
        )
        decoder_out = decoder_outs[0]
        decoder_out = torch.log_softmax(decoder_out, dim=-1)
        return decoder_out, ys_pad_lens

    def cal_decoder_with_predictor_chunk(self, encoder_out, sematic_embeds, cache=None):
        decoder_outs = self.decoder.forward_chunk(
            encoder_out, sematic_embeds, cache["decoder"]