update

嘉渊

2023-04-27 607073619cedf2c114e1589aa6d5953d171f33bf

 funasr/models/e2e_asr_paraformer.py

@@ -12,22 +12,25 @@
import numpy as np
from typeguard import check_argument_types

from funasr.layers.abs_normalize import AbsNormalize
from funasr.losses.label_smoothing_loss import (
    LabelSmoothingLoss,  # noqa: H301
)
from funasr.models.ctc import CTC
from funasr.models.decoder.abs_decoder import AbsDecoder
from funasr.models.e2e_asr_common import ErrorCalculator
from funasr.models.encoder.abs_encoder import AbsEncoder
from funasr.models.frontend.abs_frontend import AbsFrontend
from funasr.models.postencoder.abs_postencoder import AbsPostEncoder
from funasr.models.predictor.cif import mae_loss
from funasr.models.preencoder.abs_preencoder import AbsPreEncoder
from funasr.models.base_model import FunASRModel
from funasr.models.specaug.abs_specaug import AbsSpecAug
from funasr.modules.add_sos_eos import add_sos_eos
from funasr.modules.nets_utils import make_pad_mask, pad_list
from funasr.modules.nets_utils import th_accuracy
from funasr.torch_utils.device_funcs import force_gatherable
from funasr.models.base_model import FunASRModel
from funasr.models.predictor.cif import CifPredictorV3


if LooseVersion(torch.__version__) >= LooseVersion("1.6.0"):
    from torch.cuda.amp import autocast
@@ -40,7 +43,7 @@

class Paraformer(FunASRModel):
    """
    Author: Speech Lab, Alibaba Group, China
    Author: Speech Lab of DAMO Academy, Alibaba Group
    Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition
    https://arxiv.org/abs/2206.08317
    """
@@ -49,10 +52,12 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            encoder: torch.nn.Module,
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            preencoder: Optional[AbsPreEncoder],
            encoder: AbsEncoder,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
            ctc_weight: float = 0.5,
@@ -92,7 +97,16 @@
        self.frontend = frontend
        self.specaug = specaug
        self.normalize = normalize
        self.preencoder = preencoder
        self.postencoder = postencoder
        self.encoder = encoder

        if not hasattr(self.encoder, "interctc_use_conditioning"):
            self.encoder.interctc_use_conditioning = False
        if self.encoder.interctc_use_conditioning:
            self.encoder.conditioning_layer = torch.nn.Linear(
                vocab_size, self.encoder.output_size()
            )

        self.error_calculator = None

@@ -138,7 +152,6 @@
            text_lengths: torch.Tensor,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Frontend + Encoder + Decoder + Calc loss

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
@@ -161,7 +174,9 @@

        # 1. Encoder
        encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)
        intermediate_outs = None
        if isinstance(encoder_out, tuple):
            intermediate_outs = encoder_out[1]
            encoder_out = encoder_out[0]

        loss_att, acc_att, cer_att, wer_att = None, None, None, None
@@ -178,6 +193,30 @@
            # Collect CTC branch stats
            stats["loss_ctc"] = loss_ctc.detach() if loss_ctc is not None else None
            stats["cer_ctc"] = cer_ctc

        # Intermediate CTC (optional)
        loss_interctc = 0.0
        if self.interctc_weight != 0.0 and intermediate_outs is not None:
            for layer_idx, intermediate_out in intermediate_outs:
                # we assume intermediate_out has the same length & padding
                # as those of encoder_out
                loss_ic, cer_ic = self._calc_ctc_loss(
                    intermediate_out, encoder_out_lens, text, text_lengths
                )
                loss_interctc = loss_interctc + loss_ic

                # Collect Intermedaite CTC stats
                stats["loss_interctc_layer{}".format(layer_idx)] = (
                    loss_ic.detach() if loss_ic is not None else None
                )
                stats["cer_interctc_layer{}".format(layer_idx)] = cer_ic

            loss_interctc = loss_interctc / len(intermediate_outs)

            # calculate whole encoder loss
            loss_ctc = (
                               1 - self.interctc_weight
                       ) * loss_ctc + self.interctc_weight * loss_interctc

        # 2b. Attention decoder branch
        if self.ctc_weight != 1.0:
@@ -229,7 +268,6 @@
            self, speech: torch.Tensor, speech_lengths: torch.Tensor
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Frontend + Encoder. Note that this method is used by asr_inference.py

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
@@ -246,8 +284,29 @@
            if self.normalize is not None:
                feats, feats_lengths = self.normalize(feats, feats_lengths)

        # 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
        encoder_out, encoder_out_lens, _ = self.encoder(feats, feats_lengths)
        # feats: (Batch, Length, Dim)
        # -> encoder_out: (Batch, Length2, Dim2)
        if self.encoder.interctc_use_conditioning:
            encoder_out, encoder_out_lens, _ = self.encoder(
                feats, feats_lengths, ctc=self.ctc
            )
        else:
            encoder_out, encoder_out_lens, _ = self.encoder(feats, feats_lengths)
        intermediate_outs = None
        if isinstance(encoder_out, tuple):
            intermediate_outs = encoder_out[1]
            encoder_out = encoder_out[0]

        # Post-encoder, e.g. NLU
        if self.postencoder is not None:
            encoder_out, encoder_out_lens = self.postencoder(
                encoder_out, encoder_out_lens
            )

        assert encoder_out.size(0) == speech.size(0), (
            encoder_out.size(),
@@ -258,45 +317,18 @@
            encoder_out_lens.max(),
        )

        if intermediate_outs is not None:
            return (encoder_out, intermediate_outs), encoder_out_lens

        return encoder_out, encoder_out_lens

    def encode_chunk(
            self, speech: torch.Tensor, speech_lengths: torch.Tensor, cache: dict = None
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Frontend + Encoder. Note that this method is used by asr_inference.py

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
        """
        with autocast(False):
            # 1. Extract feats
            feats, feats_lengths = self._extract_feats(speech, speech_lengths)

            # 2. Data augmentation
            if self.specaug is not None and self.training:
                feats, feats_lengths = self.specaug(feats, feats_lengths)

            # 3. Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
            if self.normalize is not None:
                feats, feats_lengths = self.normalize(feats, feats_lengths)

        # 4. Forward encoder
        encoder_out, encoder_out_lens, _ = self.encoder.forward_chunk(feats, feats_lengths, cache=cache["encoder"])

        return encoder_out, torch.tensor([encoder_out.size(1)])

    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)
        pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index = self.predictor(encoder_out, None, encoder_out_mask,
                                                                                  ignore_id=self.ignore_id)
        return pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index

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

        pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index = self.predictor.forward_chunk(encoder_out, cache["encoder"])
        pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index = self.predictor(encoder_out, None,
                                                                                       encoder_out_mask,
                                                                                       ignore_id=self.ignore_id)
        return pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index

    def cal_decoder_with_predictor(self, encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens):
@@ -307,14 +339,6 @@
        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"]
        )
        decoder_out = decoder_outs
        decoder_out = torch.log_softmax(decoder_out, dim=-1)
        return decoder_out

    def _extract_feats(
            self, speech: torch.Tensor, speech_lengths: torch.Tensor
@@ -342,9 +366,7 @@
            ys_pad_lens: torch.Tensor,
    ) -> torch.Tensor:
        """Compute negative log likelihood(nll) from transformer-decoder

        Normally, this function is called in batchify_nll.

        Args:
                encoder_out: (Batch, Length, Dim)
                encoder_out_lens: (Batch,)
@@ -381,7 +403,6 @@
            batch_size: int = 100,
    ):
        """Compute negative log likelihood(nll) from transformer-decoder

        To avoid OOM, this fuction seperate the input into batches.
        Then call nll for each batch and combine and return results.
        Args:
@@ -521,9 +542,186 @@
        return loss_ctc, cer_ctc


class ParaformerBert(Paraformer):
class ParaformerOnline(Paraformer):
    """
    Author: Speech Lab, Alibaba Group, China
    Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition
    https://arxiv.org/abs/2206.08317
    """

    def __init__(
            self, *args, **kwargs,
    ):
        super().__init__(*args, **kwargs)

    def forward(
            self,
            speech: torch.Tensor,
            speech_lengths: torch.Tensor,
            text: torch.Tensor,
            text_lengths: torch.Tensor,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Frontend + Encoder + Decoder + Calc loss
        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
                text: (Batch, Length)
                text_lengths: (Batch,)
        """
        assert text_lengths.dim() == 1, text_lengths.shape
        # Check that batch_size is unified
        assert (
                speech.shape[0]
                == speech_lengths.shape[0]
                == text.shape[0]
                == text_lengths.shape[0]
        ), (speech.shape, speech_lengths.shape, text.shape, text_lengths.shape)
        batch_size = speech.shape[0]
        self.step_cur += 1
        # for data-parallel
        text = text[:, : text_lengths.max()]
        speech = speech[:, :speech_lengths.max()]

        # 1. Encoder
        encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)
        intermediate_outs = None
        if isinstance(encoder_out, tuple):
            intermediate_outs = encoder_out[1]
            encoder_out = encoder_out[0]

        loss_att, acc_att, cer_att, wer_att = None, None, None, None
        loss_ctc, cer_ctc = None, None
        loss_pre = None
        stats = dict()

        # 1. CTC branch
        if self.ctc_weight != 0.0:
            loss_ctc, cer_ctc = self._calc_ctc_loss(
                encoder_out, encoder_out_lens, text, text_lengths
            )

            # Collect CTC branch stats
            stats["loss_ctc"] = loss_ctc.detach() if loss_ctc is not None else None
            stats["cer_ctc"] = cer_ctc

        # Intermediate CTC (optional)
        loss_interctc = 0.0
        if self.interctc_weight != 0.0 and intermediate_outs is not None:
            for layer_idx, intermediate_out in intermediate_outs:
                # we assume intermediate_out has the same length & padding
                # as those of encoder_out
                loss_ic, cer_ic = self._calc_ctc_loss(
                    intermediate_out, encoder_out_lens, text, text_lengths
                )
                loss_interctc = loss_interctc + loss_ic

                # Collect Intermedaite CTC stats
                stats["loss_interctc_layer{}".format(layer_idx)] = (
                    loss_ic.detach() if loss_ic is not None else None
                )
                stats["cer_interctc_layer{}".format(layer_idx)] = cer_ic

            loss_interctc = loss_interctc / len(intermediate_outs)

            # calculate whole encoder loss
            loss_ctc = (
                               1 - self.interctc_weight
                       ) * loss_ctc + self.interctc_weight * loss_interctc

        # 2b. Attention decoder branch
        if self.ctc_weight != 1.0:
            loss_att, acc_att, cer_att, wer_att, loss_pre = self._calc_att_loss(
                encoder_out, encoder_out_lens, text, text_lengths
            )

        # 3. CTC-Att loss definition
        if self.ctc_weight == 0.0:
            loss = loss_att + loss_pre * self.predictor_weight
        elif self.ctc_weight == 1.0:
            loss = loss_ctc
        else:
            loss = self.ctc_weight * loss_ctc + (1 - self.ctc_weight) * loss_att + loss_pre * self.predictor_weight

        # Collect Attn branch stats
        stats["loss_att"] = loss_att.detach() if loss_att is not None else None
        stats["acc"] = acc_att
        stats["cer"] = cer_att
        stats["wer"] = wer_att
        stats["loss_pre"] = loss_pre.detach().cpu() if loss_pre is not None else None

        stats["loss"] = torch.clone(loss.detach())

        # force_gatherable: to-device and to-tensor if scalar for DataParallel
        loss, stats, weight = force_gatherable((loss, stats, batch_size), loss.device)
        return loss, stats, weight

    def encode_chunk(
            self, speech: torch.Tensor, speech_lengths: torch.Tensor, cache: dict = None
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Frontend + Encoder. Note that this method is used by asr_inference.py
        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
        """
        with autocast(False):
            # 1. Extract feats
            feats, feats_lengths = self._extract_feats(speech, speech_lengths)

            # 2. Data augmentation
            if self.specaug is not None and self.training:
                feats, feats_lengths = self.specaug(feats, feats_lengths)

            # 3. Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
            if self.normalize is not None:
                feats, feats_lengths = self.normalize(feats, feats_lengths)

        # 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)
        if self.encoder.interctc_use_conditioning:
            encoder_out, encoder_out_lens, _ = self.encoder.forward_chunk(
                feats, feats_lengths, cache=cache["encoder"], ctc=self.ctc
            )
        else:
            encoder_out, encoder_out_lens, _ = self.encoder.forward_chunk(feats, feats_lengths, cache=cache["encoder"])
        intermediate_outs = None
        if isinstance(encoder_out, tuple):
            intermediate_outs = encoder_out[1]
            encoder_out = encoder_out[0]

        # Post-encoder, e.g. NLU
        if self.postencoder is not None:
            encoder_out, encoder_out_lens = self.postencoder(
                encoder_out, encoder_out_lens
            )

        if intermediate_outs is not None:
            return (encoder_out, intermediate_outs), encoder_out_lens

        return encoder_out, torch.tensor([encoder_out.size(1)])

    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_chunk(self, encoder_out, sematic_embeds, cache=None):
        decoder_outs = self.decoder.forward_chunk(
            encoder_out, sematic_embeds, cache["decoder"]
        )
        decoder_out = decoder_outs
        decoder_out = torch.log_softmax(decoder_out, dim=-1)
        return decoder_out


class ParaformerBert(Paraformer):
    """
    Author: Speech Lab of DAMO Academy, Alibaba Group
    Paraformer2: advanced paraformer with LFMMI and bert for non-autoregressive end-to-end speech recognition
    """

@@ -531,11 +729,11 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            preencoder: Optional[AbsPreEncoder],
            encoder: torch.nn.Module,
            encoder: AbsEncoder,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
@@ -690,7 +888,6 @@
            embed_lengths: torch.Tensor = None,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Frontend + Encoder + Decoder + Calc loss

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
@@ -799,74 +996,73 @@


class BiCifParaformer(Paraformer):

    """
    Paraformer model with an extra cif predictor
    to conduct accurate timestamp prediction
    """

    def __init__(
        self,
        vocab_size: int,
        token_list: Union[Tuple[str, ...], List[str]],
        frontend: Optional[torch.nn.Module],
        specaug: Optional[torch.nn.Module],
        normalize: Optional[torch.nn.Module],
        preencoder: Optional[AbsPreEncoder],
        encoder: torch.nn.Module,
        postencoder: Optional[AbsPostEncoder],
        decoder: AbsDecoder,
        ctc: CTC,
        ctc_weight: float = 0.5,
        interctc_weight: float = 0.0,
        ignore_id: int = -1,
        blank_id: int = 0,
        sos: int = 1,
        eos: int = 2,
        lsm_weight: float = 0.0,
        length_normalized_loss: bool = False,
        report_cer: bool = True,
        report_wer: bool = True,
        sym_space: str = "<space>",
        sym_blank: str = "<blank>",
        extract_feats_in_collect_stats: bool = True,
        predictor = None,
        predictor_weight: float = 0.0,
        predictor_bias: int = 0,
        sampling_ratio: float = 0.2,
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            preencoder: Optional[AbsPreEncoder],
            encoder: AbsEncoder,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
            ctc_weight: float = 0.5,
            interctc_weight: float = 0.0,
            ignore_id: int = -1,
            blank_id: int = 0,
            sos: int = 1,
            eos: int = 2,
            lsm_weight: float = 0.0,
            length_normalized_loss: bool = False,
            report_cer: bool = True,
            report_wer: bool = True,
            sym_space: str = "<space>",
            sym_blank: str = "<blank>",
            extract_feats_in_collect_stats: bool = True,
            predictor=None,
            predictor_weight: float = 0.0,
            predictor_bias: int = 0,
            sampling_ratio: float = 0.2,
    ):
        assert check_argument_types()
        assert 0.0 <= ctc_weight <= 1.0, ctc_weight
        assert 0.0 <= interctc_weight < 1.0, interctc_weight

        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,
            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,
        )
        assert isinstance(self.predictor, CifPredictorV3), "BiCifParaformer should use CIFPredictorV3"

@@ -888,21 +1084,77 @@
        loss_pre2 = self.criterion_pre(ys_pad_lens.type_as(pre_token_length2), pre_token_length2)

        return loss_pre2
    

    def _calc_att_loss(
            self,
            encoder_out: torch.Tensor,
            encoder_out_lens: torch.Tensor,
            ys_pad: torch.Tensor,
            ys_pad_lens: torch.Tensor,
    ):
        encoder_out_mask = (~make_pad_mask(encoder_out_lens, maxlen=encoder_out.size(1))[:, None, :]).to(
            encoder_out.device)
        if self.predictor_bias == 1:
            _, ys_pad = add_sos_eos(ys_pad, self.sos, self.eos, self.ignore_id)
            ys_pad_lens = ys_pad_lens + self.predictor_bias
        pre_acoustic_embeds, pre_token_length, _, pre_peak_index, _ = self.predictor(encoder_out, ys_pad,
                                                                                     encoder_out_mask,
                                                                                     ignore_id=self.ignore_id)

        # 0. sampler
        decoder_out_1st = None
        if self.sampling_ratio > 0.0:
            if self.step_cur < 2:
                logging.info("enable sampler in paraformer, sampling_ratio: {}".format(self.sampling_ratio))
            sematic_embeds, decoder_out_1st = self.sampler(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens,
                                                           pre_acoustic_embeds)
        else:
            if self.step_cur < 2:
                logging.info("disable sampler in paraformer, sampling_ratio: {}".format(self.sampling_ratio))
            sematic_embeds = pre_acoustic_embeds

        # 1. Forward decoder
        decoder_outs = self.decoder(
            encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens
        )
        decoder_out, _ = decoder_outs[0], decoder_outs[1]

        if decoder_out_1st is None:
            decoder_out_1st = decoder_out
        # 2. Compute attention loss
        loss_att = self.criterion_att(decoder_out, ys_pad)
        acc_att = th_accuracy(
            decoder_out_1st.view(-1, self.vocab_size),
            ys_pad,
            ignore_label=self.ignore_id,
        )
        loss_pre = self.criterion_pre(ys_pad_lens.type_as(pre_token_length), pre_token_length)

        # Compute cer/wer using attention-decoder
        if self.training or self.error_calculator is None:
            cer_att, wer_att = None, None
        else:
            ys_hat = decoder_out_1st.argmax(dim=-1)
            cer_att, wer_att = self.error_calculator(ys_hat.cpu(), ys_pad.cpu())

        return loss_att, acc_att, cer_att, wer_att, loss_pre

    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)
        pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index, pre_token_length2 = self.predictor(encoder_out, None, encoder_out_mask,
                                                                                  ignore_id=self.ignore_id)
        pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index, pre_token_length2 = self.predictor(encoder_out,
                                                                                                          None,
                                                                                                          encoder_out_mask,
                                                                                                          ignore_id=self.ignore_id)
        return pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index
    

    def calc_predictor_timestamp(self, encoder_out, encoder_out_lens, token_num):
        encoder_out_mask = (~make_pad_mask(encoder_out_lens, maxlen=encoder_out.size(1))[:, None, :]).to(
            encoder_out.device)
        ds_alphas, ds_cif_peak, us_alphas, us_peaks = self.predictor.get_upsample_timestamp(encoder_out,
                                                                                               encoder_out_mask,
                                                                                               token_num)
                                                                                            encoder_out_mask,
                                                                                            token_num)
        return ds_alphas, ds_cif_peak, us_alphas, us_peaks

    def forward(
@@ -913,7 +1165,6 @@
            text_lengths: torch.Tensor,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Frontend + Encoder + Decoder + Calc loss

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
@@ -996,7 +1247,8 @@
        elif self.ctc_weight == 1.0:
            loss = loss_ctc
        else:
            loss = self.ctc_weight * loss_ctc + (1 - self.ctc_weight) * loss_att + loss_pre * self.predictor_weight + loss_pre2 * self.predictor_weight * 0.5
            loss = self.ctc_weight * loss_ctc + (
                        1 - self.ctc_weight) * loss_att + loss_pre * self.predictor_weight + loss_pre2 * self.predictor_weight * 0.5

        # Collect Attn branch stats
        stats["loss_att"] = loss_att.detach() if loss_att is not None else None
@@ -1022,11 +1274,11 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            preencoder: Optional[AbsPreEncoder],
            encoder: torch.nn.Module,
            encoder: AbsEncoder,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
@@ -1120,7 +1372,6 @@
            text_lengths: torch.Tensor,
    ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
        """Frontend + Encoder + Decoder + Calc loss

        Args:
                speech: (Batch, Length, ...)
                speech_lengths: (Batch, )
@@ -1504,4 +1755,4 @@
                    "torch tensor: {}, {}, loading from tf tensor: {}, {}".format(name, data_tf.size(), name_tf,
                                                                                  var_dict_tf[name_tf].shape))

        return var_dict_torch_update
        return var_dict_torch_update