Merge pull request #1825 from modelscope/dev_libt

Shi Xian

2024-06-18 6c467e6f0abfc6d20d0621fbbf67b4dbd81776cc

 funasr/models/lcbnet/model.py

@@ -15,13 +15,15 @@
from funasr.models.ctc.ctc import CTC
from funasr.models.transformer.utils.add_sos_eos import add_sos_eos
from funasr.metrics.compute_acc import th_accuracy

# from funasr.models.e2e_asr_common import ErrorCalculator
from funasr.train_utils.device_funcs import force_gatherable
from funasr.utils.load_utils import load_audio_text_image_video, extract_fbank
from funasr.utils import postprocess_utils
from funasr.utils.datadir_writer import DatadirWriter
from funasr.register import tables
import pdb


@tables.register("model_classes", "LCBNet")
class LCBNet(nn.Module):
    """
@@ -29,7 +31,7 @@
    LCB-NET: LONG-CONTEXT BIASING FOR AUDIO-VISUAL SPEECH RECOGNITION
    https://arxiv.org/abs/2401.06390
    """
    

    def __init__(
        self,
        specaug: str = None,
@@ -100,17 +102,15 @@
                **decoder_conf,
            )
        if ctc_weight > 0.0:
            

            if ctc_conf is None:
                ctc_conf = {}
            
            ctc = CTC(
                odim=vocab_size, encoder_output_size=encoder_output_size, **ctc_conf
            )
    

            ctc = CTC(odim=vocab_size, encoder_output_size=encoder_output_size, **ctc_conf)

        self.blank_id = blank_id
        self.sos = sos if sos is not None else vocab_size - 1
        self.eos = eos if eos is not None else vocab_size - 1
        self.sos = vocab_size - 1
        self.eos = vocab_size - 1
        self.vocab_size = vocab_size
        self.ignore_id = ignore_id
        self.ctc_weight = ctc_weight
@@ -138,7 +138,7 @@
            self.decoder = None
        else:
            self.decoder = decoder
        

        self.criterion_att = LabelSmoothingLoss(
            size=vocab_size,
            padding_idx=ignore_id,
@@ -156,14 +156,14 @@
            self.ctc = None
        else:
            self.ctc = ctc
            

        self.share_embedding = share_embedding
        if self.share_embedding:
            self.decoder.embed = None
        

        self.length_normalized_loss = length_normalized_loss
        self.beam_search = None
    

    def forward(
        self,
        speech: torch.Tensor,
@@ -179,36 +179,35 @@
                text: (Batch, Length)
                text_lengths: (Batch,)
        """
        # import pdb;
        # pdb.set_trace()

        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)
        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
        stats = dict()
        

        # decoder: 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:
@@ -219,25 +218,23 @@
                    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
        
            loss_ctc = (1 - self.interctc_weight) * loss_ctc + self.interctc_weight * loss_interctc

        # decoder: Attention decoder branch
        loss_att, acc_att, cer_att, wer_att = 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
@@ -245,25 +242,27 @@
            loss = loss_ctc
        else:
            loss = self.ctc_weight * loss_ctc + (1 - self.ctc_weight) * loss_att
        

        # 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
        

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

        # force_gatherable: to-device and to-tensor if scalar for DataParallel
        if self.length_normalized_loss:
            batch_size = int((text_lengths + 1).sum())
        loss, stats, weight = force_gatherable((loss, stats, batch_size), loss.device)
        return loss, stats, weight
    

    def encode(
        self, speech: torch.Tensor, speech_lengths: torch.Tensor, **kwargs,
        self,
        speech: torch.Tensor,
        speech_lengths: torch.Tensor,
        **kwargs,
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Frontend + Encoder. Note that this method is used by asr_inference.py
        Args:
@@ -272,34 +271,28 @@
                ind: int
        """
        with autocast(False):

            # Data augmentation
            if self.specaug is not None and self.training:
                speech, speech_lengths = self.specaug(speech, speech_lengths)
            
            # Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
            if self.normalize is not None:
                speech, speech_lengths = self.normalize(speech, speech_lengths)
        
        # Forward encoder
        # feats: (Batch, Length, Dim)
        # -> encoder_out: (Batch, Length2, Dim2)
        if self.encoder.interctc_use_conditioning:
            encoder_out, encoder_out_lens, _ = self.encoder(
                speech, speech_lengths, ctc=self.ctc
            )
            encoder_out, encoder_out_lens, _ = self.encoder(speech, speech_lengths, ctc=self.ctc)
        else:
            encoder_out, encoder_out_lens, _ = self.encoder(speech, speech_lengths)
        intermediate_outs = None
        if isinstance(encoder_out, tuple):
            intermediate_outs = encoder_out[1]
            encoder_out = encoder_out[0]
        

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

    def _calc_att_loss(
        self,
        encoder_out: torch.Tensor,
@@ -309,12 +302,10 @@
    ):
        ys_in_pad, ys_out_pad = add_sos_eos(ys_pad, self.sos, self.eos, self.ignore_id)
        ys_in_lens = ys_pad_lens + 1
        

        # 1. Forward decoder
        decoder_out, _ = self.decoder(
            encoder_out, encoder_out_lens, ys_in_pad, ys_in_lens
        )
        
        decoder_out, _ = self.decoder(encoder_out, encoder_out_lens, ys_in_pad, ys_in_lens)

        # 2. Compute attention loss
        loss_att = self.criterion_att(decoder_out, ys_out_pad)
        acc_att = th_accuracy(
@@ -322,16 +313,16 @@
            ys_out_pad,
            ignore_label=self.ignore_id,
        )
        

        # 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.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
    

    def _calc_ctc_loss(
        self,
        encoder_out: torch.Tensor,
@@ -341,50 +332,48 @@
    ):
        # Calc CTC loss
        loss_ctc = self.ctc(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens)
        

        # Calc CER using CTC
        cer_ctc = None
        if not self.training and self.error_calculator is not None:
            ys_hat = self.ctc.argmax(encoder_out).data
            cer_ctc = self.error_calculator(ys_hat.cpu(), ys_pad.cpu(), is_ctc=True)
        return loss_ctc, cer_ctc
    
    def init_beam_search(self,
                         **kwargs,
                         ):

    def init_beam_search(
        self,
        **kwargs,
    ):
        from funasr.models.transformer.search import BeamSearch
        from funasr.models.transformer.scorers.ctc import CTCPrefixScorer
        from funasr.models.transformer.scorers.length_bonus import LengthBonus
    

        # 1. Build ASR model
        scorers = {}
        

        if self.ctc != None:
            ctc = CTCPrefixScorer(ctc=self.ctc, eos=self.eos)
            scorers.update(
                ctc=ctc
            )
            scorers.update(ctc=ctc)
        token_list = kwargs.get("token_list")
        scorers.update(
            decoder=self.decoder,
            length_bonus=LengthBonus(len(token_list)),
        )

        
        # 3. Build ngram model
        # ngram is not supported now
        ngram = None
        scorers["ngram"] = ngram
        

        weights = dict(
            decoder=1.0 - kwargs.get("decoding_ctc_weight", 0.5),
            ctc=kwargs.get("decoding_ctc_weight", 0.5),
            decoder=1.0 - kwargs.get("decoding_ctc_weight", 0.3),
            ctc=kwargs.get("decoding_ctc_weight", 0.3),
            lm=kwargs.get("lm_weight", 0.0),
            ngram=kwargs.get("ngram_weight", 0.0),
            length_bonus=kwargs.get("penalty", 0.0),
        )
        beam_search = BeamSearch(
            beam_size=kwargs.get("beam_size", 10),
            beam_size=kwargs.get("beam_size", 20),
            weights=weights,
            scorers=scorers,
            sos=self.sos,
@@ -395,19 +384,20 @@
        )

        self.beam_search = beam_search
        
    def inference(self,
             data_in,
             data_lengths=None,
             key: list=None,
             tokenizer=None,
             frontend=None,
             **kwargs,
             ):
        

    def inference(
        self,
        data_in,
        data_lengths=None,
        key: list = None,
        tokenizer=None,
        frontend=None,
        **kwargs,
    ):

        if kwargs.get("batch_size", 1) > 1:
            raise NotImplementedError("batch decoding is not implemented")
        

        # init beamsearch
        if self.beam_search is None:
            logging.info("enable beam_search")
@@ -415,7 +405,9 @@
            self.nbest = kwargs.get("nbest", 1)

        meta_data = {}
        if isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank":  # fbank
        if (
            isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank"
        ):  # fbank
            speech, speech_lengths = data_in, data_lengths
            if len(speech.shape) < 3:
                speech = speech[None, :, :]
@@ -424,21 +416,27 @@
        else:
            # extract fbank feats
            time1 = time.perf_counter()
            pdb.set_trace()
            sample_list = load_audio_text_image_video(data_in, fs=frontend.fs, audio_fs=kwargs.get("fs", 16000),
                                                            data_type=kwargs.get("data_type", "sound"),
                                                            tokenizer=tokenizer)
            sample_list = load_audio_text_image_video(
                data_in,
                fs=frontend.fs,
                audio_fs=kwargs.get("fs", 16000),
                data_type=kwargs.get("data_type", "sound"),
                tokenizer=tokenizer,
            )
            time2 = time.perf_counter()
            meta_data["load_data"] = f"{time2 - time1:0.3f}"
            audio_sample_list = sample_list[0]
            ocr_sample_list = sample_list[1]
            pdb.set_trace()
            speech, speech_lengths = extract_fbank(audio_sample_list, data_type=kwargs.get("data_type", "sound"),
                                                   frontend=frontend)
            pdb.set_trace()
            if len(sample_list) > 1:
                ocr_sample_list = sample_list[1]
            else:
                ocr_sample_list = [[294, 0]]
            speech, speech_lengths = extract_fbank(
                audio_sample_list, data_type=kwargs.get("data_type", "sound"), frontend=frontend
            )
            time3 = time.perf_counter()
            meta_data["extract_feat"] = f"{time3 - time2:0.3f}"
            meta_data["batch_data_time"] = speech_lengths.sum().item() * frontend.frame_shift * frontend.lfr_n / 1000
            frame_shift = 10
            meta_data["batch_data_time"] = speech_lengths.sum().item() * frame_shift / 1000

        speech = speech.to(device=kwargs["device"])
        speech_lengths = speech_lengths.to(device=kwargs["device"])
@@ -446,14 +444,23 @@
        encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)
        if isinstance(encoder_out, tuple):
            encoder_out = encoder_out[0]
        

        ocr_list_new = [[x + 1 if x != 0 else x for x in sublist] for sublist in ocr_sample_list]
        ocr = torch.tensor(ocr_list_new).to(device=kwargs["device"])
        ocr_lengths = ocr.new_full([1], dtype=torch.long, fill_value=ocr.size(1)).to(
            device=kwargs["device"]
        )
        ocr, ocr_lens, _ = self.text_encoder(ocr, ocr_lengths)
        fusion_out, _, _, _ = self.fusion_encoder(encoder_out, None, ocr, None)
        encoder_out = encoder_out + fusion_out
        # c. Passed the encoder result and the beam search
        nbest_hyps = self.beam_search(
            x=encoder_out[0], maxlenratio=kwargs.get("maxlenratio", 0.0), minlenratio=kwargs.get("minlenratio", 0.0)
            x=encoder_out[0],
            maxlenratio=kwargs.get("maxlenratio", 0.0),
            minlenratio=kwargs.get("minlenratio", 0.0),
        )
        
        nbest_hyps = nbest_hyps[: self.nbest]

        nbest_hyps = nbest_hyps[: self.nbest]

        results = []
        b, n, d = encoder_out.size()
@@ -465,28 +472,31 @@
                    if not hasattr(self, "writer"):
                        self.writer = DatadirWriter(kwargs.get("output_dir"))
                    ibest_writer = self.writer[f"{nbest_idx + 1}best_recog"]
                    

                # remove sos/eos and get results
                last_pos = -1
                if isinstance(hyp.yseq, list):
                    token_int = hyp.yseq[1:last_pos]
                else:
                    token_int = hyp.yseq[1:last_pos].tolist()
                    

                # remove blank symbol id, which is assumed to be 0
                token_int = list(filter(lambda x: x != self.eos and x != self.sos and x != self.blank_id, token_int))
                
                token_int = list(
                    filter(
                        lambda x: x != self.eos and x != self.sos and x != self.blank_id, token_int
                    )
                )

                # Change integer-ids to tokens
                token = tokenizer.ids2tokens(token_int)
                text = tokenizer.tokens2text(token)
                

                text_postprocessed, _ = postprocess_utils.sentence_postprocess(token)
                result_i = {"key": key[i], "token": token, "text": text_postprocessed}
                results.append(result_i)
                

                if ibest_writer is not None:
                    ibest_writer["token"][key[i]] = " ".join(token)
                    ibest_writer["text"][key[i]] = text_postprocessed
        
        return results, meta_data

        return results, meta_data