Merge branch 'main' of github.com:alibaba-damo-academy/FunASR add

游雁

2023-09-13 33d3d2084403fd34b79c835d2f2fe04f6cd8f738

 funasr/models/e2e_asr_transducer.py

@@ -6,8 +6,9 @@

import torch
from packaging.version import parse as V
from typeguard import check_argument_types

from funasr.losses.label_smoothing_loss import (
    LabelSmoothingLoss,  # noqa: H301
)
from funasr.models.frontend.abs_frontend import AbsFrontend
from funasr.models.specaug.abs_specaug import AbsSpecAug
from funasr.models.decoder.rnnt_decoder import RNNTDecoder
@@ -15,6 +16,8 @@
from funasr.models.encoder.abs_encoder import AbsEncoder
from funasr.models.joint_net.joint_network import JointNetwork
from funasr.modules.nets_utils import get_transducer_task_io
from funasr.modules.nets_utils import th_accuracy
from funasr.modules.add_sos_eos import add_sos_eos
from funasr.layers.abs_normalize import AbsNormalize
from funasr.torch_utils.device_funcs import force_gatherable
from funasr.models.base_model import FunASRModel
@@ -81,8 +84,6 @@
    ) -> None:
        """Construct an ESPnetASRTransducerModel object."""
        super().__init__()

        assert check_argument_types()

        # The following labels ID are reserved: 0 (blank) and vocab_size - 1 (sos/eos)
        self.blank_id = 0
@@ -352,11 +353,6 @@
        """
        if self.criterion_transducer is None:
            try:
                # from warprnnt_pytorch import RNNTLoss
           # self.criterion_transducer = RNNTLoss(
                    # reduction="mean",
                    # fastemit_lambda=self.fastemit_lambda,
                # )
                from warp_rnnt import rnnt_loss as RNNTLoss
                self.criterion_transducer = RNNTLoss

@@ -367,12 +363,6 @@
                )
                exit(1)

        # loss_transducer = self.criterion_transducer(
        #     joint_out,
        #     target,
        #     t_len,
        #     u_len,
        # )
        log_probs = torch.log_softmax(joint_out, dim=-1)

        loss_transducer = self.criterion_transducer(
@@ -542,8 +532,6 @@
        """Construct an ESPnetASRTransducerModel object."""
        super().__init__()

        assert check_argument_types()

        # The following labels ID are reserved: 0 (blank) and vocab_size - 1 (sos/eos)
        self.blank_id = 0

@@ -638,7 +626,6 @@

        batch_size = speech.shape[0]
        text = text[:, : text_lengths.max()]
        #print(speech.shape)
        # 1. Encoder
        encoder_out, encoder_out_chunk, encoder_out_lens = self.encode(speech, speech_lengths)

@@ -709,8 +696,8 @@
            loss_lm = self._calc_lm_loss(decoder_out, target)

        loss_trans = loss_trans_utt + loss_trans_chunk
        loss_ctc = loss_ctc + loss_ctc_chunk 
        loss_ctc = loss_att + loss_att_chunk
        loss_ctc = loss_ctc + loss_ctc_chunk
        loss_att = loss_att + loss_att_chunk

        loss = (
            self.transducer_weight * loss_trans
@@ -855,11 +842,6 @@
        """
        if self.criterion_transducer is None:
            try:
                # from warprnnt_pytorch import RNNTLoss
            # self.criterion_transducer = RNNTLoss(
                    # reduction="mean",
                    # fastemit_lambda=self.fastemit_lambda,
                # )
                from warp_rnnt import rnnt_loss as RNNTLoss
                self.criterion_transducer = RNNTLoss

@@ -870,12 +852,6 @@
                )
                exit(1)

        # loss_transducer = self.criterion_transducer(
        #     joint_out,
        #     target,
        #     t_len,
        #     u_len,
        # )
        log_probs = torch.log_softmax(joint_out, dim=-1)

        loss_transducer = self.criterion_transducer(