update

嘉渊

2023-04-24 6427c834dfd97b1f05c6659cdc7ccf010bf82fe1

 funasr/models/e2e_asr_paraformer.py

@@ -12,24 +12,20 @@
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.specaug.abs_specaug import AbsSpecAug
from funasr.models.base_model import FunASRModel
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.train.abs_espnet_model import AbsESPnetModel
from funasr.models.predictor.cif import CifPredictorV3


@@ -42,7 +38,7 @@
        yield


class Paraformer(AbsESPnetModel):
class Paraformer(FunASRModel):
    """
    Author: Speech Lab, Alibaba Group, China
    Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition
@@ -53,11 +49,11 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            preencoder: Optional[AbsPreEncoder],
            encoder: AbsEncoder,
            encoder: torch.nn.Module,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
@@ -325,12 +321,67 @@

        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)

        # 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(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"])
        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):
@@ -341,6 +392,14 @@
        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
@@ -557,11 +616,11 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            preencoder: Optional[AbsPreEncoder],
            encoder: AbsEncoder,
            encoder: torch.nn.Module,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
@@ -835,11 +894,11 @@
        self,
        vocab_size: int,
        token_list: Union[Tuple[str, ...], List[str]],
        frontend: Optional[AbsFrontend],
        specaug: Optional[AbsSpecAug],
        normalize: Optional[AbsNormalize],
        frontend: Optional[torch.nn.Module],
        specaug: Optional[torch.nn.Module],
        normalize: Optional[torch.nn.Module],
        preencoder: Optional[AbsPreEncoder],
        encoder: AbsEncoder,
        encoder: torch.nn.Module,
        postencoder: Optional[AbsPostEncoder],
        decoder: AbsDecoder,
        ctc: CTC,
@@ -926,10 +985,10 @@
    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_cif_peak = self.predictor.get_upsample_timestamp(encoder_out,
        ds_alphas, ds_cif_peak, us_alphas, us_peaks = self.predictor.get_upsample_timestamp(encoder_out,
                                                                                               encoder_out_mask,
                                                                                               token_num)
        return ds_alphas, ds_cif_peak, us_alphas, us_cif_peak
        return ds_alphas, ds_cif_peak, us_alphas, us_peaks

    def forward(
            self,
@@ -962,21 +1021,82 @@

        # 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
            )

        loss_pre2 = self._calc_pre2_loss(
            encoder_out, encoder_out_lens, text, text_lengths
        )

        loss = loss_pre2
        # 3. CTC-Att loss definition
        if self.ctc_weight == 0.0:
            loss = loss_att + loss_pre * self.predictor_weight + loss_pre2 * self.predictor_weight * 0.5
        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

        # 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_pre2"] = loss_pre2.detach().cpu()

        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


class ContextualParaformer(Paraformer):
    """
@@ -987,11 +1107,11 @@
            self,
            vocab_size: int,
            token_list: Union[Tuple[str, ...], List[str]],
            frontend: Optional[AbsFrontend],
            specaug: Optional[AbsSpecAug],
            normalize: Optional[AbsNormalize],
            frontend: Optional[torch.nn.Module],
            specaug: Optional[torch.nn.Module],
            normalize: Optional[torch.nn.Module],
            preencoder: Optional[AbsPreEncoder],
            encoder: AbsEncoder,
            encoder: torch.nn.Module,
            postencoder: Optional[AbsPostEncoder],
            decoder: AbsDecoder,
            ctc: CTC,
@@ -1021,6 +1141,7 @@
            inner_dim: int = 256,
            bias_encoder_type: str = 'lstm',
            label_bracket: bool = False,
            use_decoder_embedding: bool = False,
    ):
        assert check_argument_types()
        assert 0.0 <= ctc_weight <= 1.0, ctc_weight
@@ -1074,6 +1195,7 @@
            self.hotword_buffer = None
            self.length_record = []
            self.current_buffer_length = 0
        self.use_decoder_embedding = use_decoder_embedding

    def forward(
            self,
@@ -1215,7 +1337,10 @@
                    hw_list.append(hw_tokens)
        # padding
        hw_list_pad = pad_list(hw_list, 0)
        hw_embed = self.decoder.embed(hw_list_pad)
        if self.use_decoder_embedding:
            hw_embed = self.decoder.embed(hw_list_pad)
        else:
            hw_embed = self.bias_embed(hw_list_pad)
        hw_embed, (_, _) = self.bias_encoder(hw_embed)
        _ind = np.arange(0, len(hw_list)).tolist()
        # update self.hotword_buffer, throw a part if oversize
@@ -1331,13 +1456,19 @@
            # default hotword list
            hw_list = [torch.Tensor([self.sos]).long().to(encoder_out.device)]  # empty hotword list
            hw_list_pad = pad_list(hw_list, 0)
            hw_embed = self.bias_embed(hw_list_pad)
            if self.use_decoder_embedding:
                hw_embed = self.decoder.embed(hw_list_pad)
            else:
                hw_embed = self.bias_embed(hw_list_pad)
            _, (h_n, _) = self.bias_encoder(hw_embed)
            contextual_info = h_n.squeeze(0).repeat(encoder_out.shape[0], 1, 1)
        else:
            hw_lengths = [len(i) for i in hw_list]
            hw_list_pad = pad_list([torch.Tensor(i).long() for i in hw_list], 0).to(encoder_out.device)
            hw_embed = self.bias_embed(hw_list_pad)
            if self.use_decoder_embedding:
                hw_embed = self.decoder.embed(hw_list_pad)
            else:
                hw_embed = self.bias_embed(hw_list_pad)
            hw_embed = torch.nn.utils.rnn.pack_padded_sequence(hw_embed, hw_lengths, batch_first=True,
                                                               enforce_sorted=False)
            _, (h_n, _) = self.bias_encoder(hw_embed)
@@ -1458,4 +1589,4 @@
                    "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