From 33d3d2084403fd34b79c835d2f2fe04f6cd8f738 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期三, 13 九月 2023 09:33:54 +0800
Subject: [PATCH] Merge branch 'main' of github.com:alibaba-damo-academy/FunASR add
---
funasr/models/e2e_asr_paraformer.py | 2800 ++++++++++++++++++++++++++++++++++++++++++++--------------
1 files changed, 2,114 insertions(+), 686 deletions(-)
diff --git a/funasr/models/e2e_asr_paraformer.py b/funasr/models/e2e_asr_paraformer.py
index 5ea28f3..e157454 100644
--- a/funasr/models/e2e_asr_paraformer.py
+++ b/funasr/models/e2e_asr_paraformer.py
@@ -8,11 +8,12 @@
from typing import Union
import torch
-from typeguard import check_argument_types
+import random
+import numpy as np
from funasr.layers.abs_normalize import AbsNormalize
from funasr.losses.label_smoothing_loss import (
- LabelSmoothingLoss, # noqa: H301
+ LabelSmoothingLoss, # noqa: H301
)
from funasr.models.ctc import CTC
from funasr.models.decoder.abs_decoder import AbsDecoder
@@ -20,801 +21,2228 @@
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.modules.add_sos_eos import add_sos_eos
-from funasr.modules.nets_utils import make_pad_mask
+from funasr.modules.nets_utils import make_pad_mask, pad_list
from funasr.modules.nets_utils import th_accuracy
-from funasr.models.predictor.cif import mae_loss
from funasr.torch_utils.device_funcs import force_gatherable
-from funasr.train.abs_espnet_model import AbsESPnetModel
+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
+ from torch.cuda.amp import autocast
else:
- # Nothing to do if torch<1.6.0
- @contextmanager
- def autocast(enabled=True):
- yield
+ # Nothing to do if torch<1.6.0
+ @contextmanager
+ def autocast(enabled=True):
+ yield
-class Paraformer(AbsESPnetModel):
- """
- 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,
- 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,
+class Paraformer(FunASRModel):
+ """
+ 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
+ """
- ):
- assert check_argument_types()
- assert 0.0 <= ctc_weight <= 1.0, ctc_weight
- assert 0.0 <= interctc_weight < 1.0, interctc_weight
+ def __init__(
+ self,
+ vocab_size: int,
+ token_list: Union[Tuple[str, ...], List[str]],
+ frontend: Optional[AbsFrontend],
+ specaug: Optional[AbsSpecAug],
+ normalize: Optional[AbsNormalize],
+ encoder: AbsEncoder,
+ 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,
+ share_embedding: bool = False,
+ preencoder: Optional[AbsPreEncoder] = None,
+ postencoder: Optional[AbsPostEncoder] = None,
+ use_1st_decoder_loss: bool = False,
+ ):
+ assert 0.0 <= ctc_weight <= 1.0, ctc_weight
+ assert 0.0 <= interctc_weight < 1.0, interctc_weight
- super().__init__()
- # 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
- self.eos = vocab_size - 1 if eos is None else eos
- self.vocab_size = vocab_size
- self.ignore_id = ignore_id
- self.ctc_weight = ctc_weight
- self.interctc_weight = interctc_weight
- self.token_list = token_list.copy()
+ super().__init__()
+ # 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
+ self.eos = vocab_size - 1 if eos is None else eos
+ self.vocab_size = vocab_size
+ self.ignore_id = ignore_id
+ self.ctc_weight = ctc_weight
+ self.interctc_weight = interctc_weight
+ self.token_list = token_list.copy()
- self.frontend = frontend
- self.specaug = specaug
- self.normalize = normalize
- self.preencoder = preencoder
- self.postencoder = postencoder
- self.encoder = encoder
+ 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()
- )
+ 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
+ self.error_calculator = None
+ if ctc_weight == 1.0:
+ self.decoder = None
+ else:
+ self.decoder = decoder
- if ctc_weight == 1.0:
- self.decoder = None
- else:
- self.decoder = decoder
+ self.criterion_att = LabelSmoothingLoss(
+ size=vocab_size,
+ padding_idx=ignore_id,
+ smoothing=lsm_weight,
+ normalize_length=length_normalized_loss,
+ )
- self.criterion_att = LabelSmoothingLoss(
- size=vocab_size,
- padding_idx=ignore_id,
- smoothing=lsm_weight,
- normalize_length=length_normalized_loss,
- )
+ if report_cer or report_wer:
+ self.error_calculator = ErrorCalculator(
+ token_list, sym_space, sym_blank, report_cer, report_wer
+ )
- if report_cer or report_wer:
- self.error_calculator = ErrorCalculator(
- token_list, sym_space, sym_blank, report_cer, report_wer
- )
+ if ctc_weight == 0.0:
+ self.ctc = None
+ else:
+ self.ctc = ctc
- if ctc_weight == 0.0:
- self.ctc = None
- else:
- self.ctc = ctc
+ self.extract_feats_in_collect_stats = extract_feats_in_collect_stats
+ self.predictor = predictor
+ self.predictor_weight = predictor_weight
+ self.predictor_bias = predictor_bias
+ self.sampling_ratio = sampling_ratio
+ self.criterion_pre = mae_loss(normalize_length=length_normalized_loss)
+ self.step_cur = 0
- self.extract_feats_in_collect_stats = extract_feats_in_collect_stats
- self.predictor = predictor
- self.predictor_weight = predictor_weight
- self.predictor_bias = predictor_bias
- self.sampling_ratio = sampling_ratio
- self.criterion_pre = mae_loss(normalize_length=length_normalized_loss)
- self.step_cur = 0
+ self.share_embedding = share_embedding
+ if self.share_embedding:
+ self.decoder.embed = None
+ self.use_1st_decoder_loss = use_1st_decoder_loss
- 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
+ def forward(
+ self,
+ speech: torch.Tensor,
+ speech_lengths: torch.Tensor,
+ text: torch.Tensor,
+ text_lengths: torch.Tensor,
+ decoding_ind: int = None,
+ ) -> 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,)
+ decoding_ind: int
+ """
+ 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
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ ind = self.encoder.overlap_chunk_cls.random_choice(self.training, decoding_ind)
+ encoder_out, encoder_out_lens = self.encode(speech, speech_lengths, ind=ind)
+ else:
+ 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, pre_loss_att, acc_att, cer_att, wer_att = None, 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, pre_loss_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 + 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
+
+ if self.use_1st_decoder_loss and pre_loss_att is not None:
+ loss = loss + (1 - self.ctc_weight) * pre_loss_att
+
+ # Collect Attn branch stats
+ stats["loss_att"] = loss_att.detach() if loss_att is not None else None
+ stats["pre_loss_att"] = pre_loss_att.detach() if pre_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 collect_feats(
+ self,
+ speech: torch.Tensor,
+ speech_lengths: torch.Tensor,
+ text: torch.Tensor,
+ text_lengths: torch.Tensor,
+ ) -> Dict[str, torch.Tensor]:
+ if self.extract_feats_in_collect_stats:
+ feats, feats_lengths = self._extract_feats(speech, speech_lengths)
+ else:
+ # Generate dummy stats if extract_feats_in_collect_stats is False
+ logging.warning(
+ "Generating dummy stats for feats and feats_lengths, "
+ "because encoder_conf.extract_feats_in_collect_stats is "
+ f"{self.extract_feats_in_collect_stats}"
+ )
+ feats, feats_lengths = speech, speech_lengths
+ return {"feats": feats, "feats_lengths": feats_lengths}
+
+ def encode(
+ self, speech: torch.Tensor, speech_lengths: torch.Tensor, ind: int = 0,
+ ) -> Tuple[torch.Tensor, torch.Tensor]:
+ """Frontend + Encoder. Note that this method is used by asr_inference.py
+ Args:
+ speech: (Batch, Length, ...)
+ speech_lengths: (Batch, )
+ ind: int
+ """
+ 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:
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ encoder_out, encoder_out_lens, _ = self.encoder(
+ feats, feats_lengths, ctc=self.ctc, ind=ind
+ )
+ encoder_out, encoder_out_lens = self.encoder.overlap_chunk_cls.remove_chunk(encoder_out,
+ encoder_out_lens,
+ chunk_outs=None)
+ else:
+ encoder_out, encoder_out_lens, _ = self.encoder(
+ feats, feats_lengths, ctc=self.ctc
+ )
+ else:
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ encoder_out, encoder_out_lens, _ = self.encoder(feats, feats_lengths, ind=ind)
+ encoder_out, encoder_out_lens = self.encoder.overlap_chunk_cls.remove_chunk(encoder_out,
+ encoder_out_lens,
+ chunk_outs=None)
+ 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(),
+ speech.size(0),
+ )
+ assert encoder_out.size(1) <= encoder_out_lens.max(), (
+ encoder_out.size(),
+ 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 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 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
+ )
+ decoder_out = decoder_outs[0]
+ decoder_out = torch.log_softmax(decoder_out, dim=-1)
+ return decoder_out, ys_pad_lens
+
+ def _extract_feats(
+ self, speech: torch.Tensor, speech_lengths: torch.Tensor
+ ) -> Tuple[torch.Tensor, torch.Tensor]:
+ assert speech_lengths.dim() == 1, speech_lengths.shape
+
+ # for data-parallel
+ speech = speech[:, : speech_lengths.max()]
+ if self.frontend is not None:
+ # Frontend
+ # e.g. STFT and Feature extract
+ # data_loader may send time-domain signal in this case
+ # speech (Batch, NSamples) -> feats: (Batch, NFrames, Dim)
+ feats, feats_lengths = self.frontend(speech, speech_lengths)
+ else:
+ # No frontend and no feature extract
+ feats, feats_lengths = speech, speech_lengths
+ return feats, feats_lengths
+
+ def nll(
+ self,
+ encoder_out: torch.Tensor,
+ encoder_out_lens: torch.Tensor,
+ ys_pad: torch.Tensor,
+ 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,)
+ ys_pad: (Batch, Length)
+ ys_pad_lens: (Batch,)
+ """
+ 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
+ ) # [batch, seqlen, dim]
+ batch_size = decoder_out.size(0)
+ decoder_num_class = decoder_out.size(2)
+ # nll: negative log-likelihood
+ nll = torch.nn.functional.cross_entropy(
+ decoder_out.view(-1, decoder_num_class),
+ ys_out_pad.view(-1),
+ ignore_index=self.ignore_id,
+ reduction="none",
+ )
+ nll = nll.view(batch_size, -1)
+ nll = nll.sum(dim=1)
+ assert nll.size(0) == batch_size
+ return nll
+
+ def batchify_nll(
+ self,
+ encoder_out: torch.Tensor,
+ encoder_out_lens: torch.Tensor,
+ ys_pad: torch.Tensor,
+ ys_pad_lens: torch.Tensor,
+ 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:
+ encoder_out: (Batch, Length, Dim)
+ encoder_out_lens: (Batch,)
+ ys_pad: (Batch, Length)
+ ys_pad_lens: (Batch,)
+ batch_size: int, samples each batch contain when computing nll,
+ you may change this to avoid OOM or increase
+ GPU memory usage
+ """
+ total_num = encoder_out.size(0)
+ if total_num <= batch_size:
+ nll = self.nll(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens)
+ else:
+ nll = []
+ start_idx = 0
+ while True:
+ end_idx = min(start_idx + batch_size, total_num)
+ batch_encoder_out = encoder_out[start_idx:end_idx, :, :]
+ batch_encoder_out_lens = encoder_out_lens[start_idx:end_idx]
+ batch_ys_pad = ys_pad[start_idx:end_idx, :]
+ batch_ys_pad_lens = ys_pad_lens[start_idx:end_idx]
+ batch_nll = self.nll(
+ batch_encoder_out,
+ batch_encoder_out_lens,
+ batch_ys_pad,
+ batch_ys_pad_lens,
+ )
+ nll.append(batch_nll)
+ start_idx = end_idx
+ if start_idx == total_num:
+ break
+ nll = torch.cat(nll)
+ assert nll.size(0) == total_num
+ return nll
+
+ 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
+ pre_loss_att = None
+ if self.sampling_ratio > 0.0:
+ if self.step_cur < 2:
+ logging.info("enable sampler in paraformer, sampling_ratio: {}".format(self.sampling_ratio))
+ if self.use_1st_decoder_loss:
+ sematic_embeds, decoder_out_1st, pre_loss_att = self.sampler_with_grad(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens,
+ pre_acoustic_embeds)
+ else:
+ 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, pre_loss_att
+
+ def sampler(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds):
+
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ ys_pad_masked = ys_pad * tgt_mask[:, :, 0]
+ if self.share_embedding:
+ ys_pad_embed = self.decoder.output_layer.weight[ys_pad_masked]
+ else:
+ ys_pad_embed = self.decoder.embed(ys_pad_masked)
+ with torch.no_grad():
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens
+ )
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
+ pred_tokens = decoder_out.argmax(-1)
+ nonpad_positions = ys_pad.ne(self.ignore_id)
+ seq_lens = (nonpad_positions).sum(1)
+ same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
+ input_mask = torch.ones_like(nonpad_positions)
+ bsz, seq_len = ys_pad.size()
+ for li in range(bsz):
+ target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
+ if target_num > 0:
+ input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
+ input_mask = input_mask.eq(1)
+ input_mask = input_mask.masked_fill(~nonpad_positions, False)
+ input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
+
+ sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
+ input_mask_expand_dim, 0)
+ return sematic_embeds * tgt_mask, decoder_out * tgt_mask
+
+ def sampler_with_grad(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds):
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ ys_pad_masked = ys_pad * tgt_mask[:, :, 0]
+ if self.share_embedding:
+ ys_pad_embed = self.decoder.output_layer.weight[ys_pad_masked]
+ else:
+ ys_pad_embed = self.decoder.embed(ys_pad_masked)
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens
+ )
+ pre_loss_att = self.criterion_att(decoder_outs[0], ys_pad)
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
+ pred_tokens = decoder_out.argmax(-1)
+ nonpad_positions = ys_pad.ne(self.ignore_id)
+ seq_lens = (nonpad_positions).sum(1)
+ same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
+ input_mask = torch.ones_like(nonpad_positions)
+ bsz, seq_len = ys_pad.size()
+ for li in range(bsz):
+ target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
+ if target_num > 0:
+ input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
+ input_mask = input_mask.eq(1)
+ input_mask = input_mask.masked_fill(~nonpad_positions, False)
+ input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
+
+ sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
+ input_mask_expand_dim, 0)
+
+ return sematic_embeds * tgt_mask, decoder_out * tgt_mask, pre_loss_att
+
+ def _calc_ctc_loss(
+ self,
+ encoder_out: torch.Tensor,
+ encoder_out_lens: torch.Tensor,
+ ys_pad: torch.Tensor,
+ ys_pad_lens: torch.Tensor,
+ ):
+ # 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
+
+
+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,
+ vocab_size: int,
+ token_list: Union[Tuple[str, ...], List[str]],
+ frontend: Optional[AbsFrontend],
+ specaug: Optional[AbsSpecAug],
+ normalize: Optional[AbsNormalize],
+ encoder: AbsEncoder,
+ 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,
+ decoder_attention_chunk_type: str = 'chunk',
+ share_embedding: bool = False,
+ preencoder: Optional[AbsPreEncoder] = None,
+ postencoder: Optional[AbsPostEncoder] = None,
+ use_1st_decoder_loss: bool = False,
+ ):
+ 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,
+ )
+ # 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
+ self.eos = vocab_size - 1 if eos is None else eos
+ self.vocab_size = vocab_size
+ self.ignore_id = ignore_id
+ self.ctc_weight = ctc_weight
+ self.interctc_weight = interctc_weight
+ self.token_list = token_list.copy()
+
+ 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
+
+ if ctc_weight == 1.0:
+ self.decoder = None
+ else:
+ self.decoder = decoder
+
+ self.criterion_att = LabelSmoothingLoss(
+ size=vocab_size,
+ padding_idx=ignore_id,
+ smoothing=lsm_weight,
+ normalize_length=length_normalized_loss,
+ )
+
+ if report_cer or report_wer:
+ self.error_calculator = ErrorCalculator(
+ token_list, sym_space, sym_blank, report_cer, report_wer
+ )
+
+ if ctc_weight == 0.0:
+ self.ctc = None
+ else:
+ self.ctc = ctc
+
+ self.extract_feats_in_collect_stats = extract_feats_in_collect_stats
+ self.predictor = predictor
+ self.predictor_weight = predictor_weight
+ self.predictor_bias = predictor_bias
+ 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
+ self.decoder_attention_chunk_type = decoder_attention_chunk_type
- 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(), :]
+ self.share_embedding = share_embedding
+ if self.share_embedding:
+ self.decoder.embed = None
- # 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]
+ self.use_1st_decoder_loss = use_1st_decoder_loss
+ def forward(
+ self,
+ speech: torch.Tensor,
+ speech_lengths: torch.Tensor,
+ text: torch.Tensor,
+ text_lengths: torch.Tensor,
+ decoding_ind: int = None,
+ ) -> 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,)
+ decoding_ind: int
+ """
+ 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()]
- loss_att, acc_att, cer_att, wer_att = None, None, None, None
- loss_ctc, cer_ctc = None, None
- loss_pre = None
- stats = dict()
+ # 1. Encoder
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ ind = self.encoder.overlap_chunk_cls.random_choice(self.training, decoding_ind)
+ encoder_out, encoder_out_lens = self.encode(speech, speech_lengths, ind=ind)
+ else:
+ 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]
- # 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
- )
+ loss_att, acc_att, cer_att, wer_att = None, None, None, None
+ loss_ctc, cer_ctc = None, None
+ loss_pre = None
+ stats = dict()
- # Collect CTC branch stats
- stats["loss_ctc"] = loss_ctc.detach() if loss_ctc is not None else None
- stats["cer_ctc"] = cer_ctc
+ # 1. CTC branch
+ if self.ctc_weight != 0.0:
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ encoder_out_ctc, encoder_out_lens_ctc = self.encoder.overlap_chunk_cls.remove_chunk(encoder_out,
+ encoder_out_lens,
+ chunk_outs=None)
+ loss_ctc, cer_ctc = self._calc_ctc_loss(
+ encoder_out_ctc, encoder_out_lens_ctc, text, text_lengths
+ )
- # 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 CTC branch stats
+ stats["loss_ctc"] = loss_ctc.detach() if loss_ctc is not None else None
+ stats["cer_ctc"] = cer_ctc
- # 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
+ # 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
+ if hasattr(self.encoder, "overlap_chunk_cls"):
+ encoder_out_ctc, encoder_out_lens_ctc = \
+ self.encoder.overlap_chunk_cls.remove_chunk(
+ intermediate_out,
+ encoder_out_lens,
+ chunk_outs=None)
+ loss_ic, cer_ic = self._calc_ctc_loss(
+ encoder_out_ctc, encoder_out_lens_ctc, text, text_lengths
+ )
+ loss_interctc = loss_interctc + loss_ic
- loss_interctc = loss_interctc / len(intermediate_outs)
+ # 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
- # calculate whole encoder loss
- loss_ctc = (
- 1 - self.interctc_weight
- ) * loss_ctc + self.interctc_weight * loss_interctc
+ 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:
+ # 2b. Attention decoder branch
+ if self.ctc_weight != 1.0:
+ loss_att, acc_att, cer_att, wer_att, loss_pre, pre_loss_att = self._calc_att_predictor_loss(
+ encoder_out, encoder_out_lens, text, text_lengths
+ )
- 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
- # 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
+ if self.use_1st_decoder_loss and pre_loss_att is not None:
+ loss = loss + pre_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
- stats["loss_pre"] = loss_pre.detach().cpu() if loss_pre is not None else None
+ # Collect Attn branch stats
+ stats["loss_att"] = loss_att.detach() if loss_att is not None else None
+ stats["pre_loss_att"] = pre_loss_att.detach() if pre_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())
+ 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
+ # 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 collect_feats(
- self,
- speech: torch.Tensor,
- speech_lengths: torch.Tensor,
- text: torch.Tensor,
- text_lengths: torch.Tensor,
- ) -> Dict[str, torch.Tensor]:
- if self.extract_feats_in_collect_stats:
- feats, feats_lengths = self._extract_feats(speech, speech_lengths)
- else:
- # Generate dummy stats if extract_feats_in_collect_stats is False
- logging.warning(
- "Generating dummy stats for feats and feats_lengths, "
- "because encoder_conf.extract_feats_in_collect_stats is "
- f"{self.extract_feats_in_collect_stats}"
- )
- feats, feats_lengths = speech, speech_lengths
- return {"feats": feats, "feats_lengths": feats_lengths}
+ def encode(
+ self, speech: torch.Tensor, speech_lengths: torch.Tensor, ind: int = 0,
+ ) -> 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)
- def encode(
- 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
+ # 2. Data augmentation
+ if self.specaug is not None and self.training:
+ feats, feats_lengths = self.specaug(feats, feats_lengths)
- Args:
- speech: (Batch, Length, ...)
- speech_lengths: (Batch, )
- """
- with autocast(False):
- # 1. Extract feats
- feats, feats_lengths = self._extract_feats(speech, speech_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(
+ feats, feats_lengths, ctc=self.ctc, ind=ind
+ )
+ else:
+ encoder_out, encoder_out_lens, _ = self.encoder(feats, feats_lengths, ind=ind)
+ intermediate_outs = None
+ if isinstance(encoder_out, tuple):
+ intermediate_outs = encoder_out[1]
+ encoder_out = encoder_out[0]
- # 2. Data augmentation
- if self.specaug is not None and self.training:
- feats, feats_lengths = self.specaug(feats, feats_lengths)
+ # Post-encoder, e.g. NLU
+ if self.postencoder is not None:
+ encoder_out, encoder_out_lens = self.postencoder(
+ encoder_out, encoder_out_lens
+ )
- # 3. Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
- if self.normalize is not None:
- feats, feats_lengths = self.normalize(feats, feats_lengths)
+ assert encoder_out.size(0) == speech.size(0), (
+ encoder_out.size(),
+ speech.size(0),
+ )
+ assert encoder_out.size(1) <= encoder_out_lens.max(), (
+ encoder_out.size(),
+ encoder_out_lens.max(),
+ )
- # Pre-encoder, e.g. used for raw input data
- if self.preencoder is not None:
- feats, feats_lengths = self.preencoder(feats, feats_lengths)
+ if intermediate_outs is not None:
+ return (encoder_out, intermediate_outs), encoder_out_lens
- # 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(
- 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]
+ return encoder_out, encoder_out_lens
- # Post-encoder, e.g. NLU
- if self.postencoder is not None:
- encoder_out, encoder_out_lens = self.postencoder(
- 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)
- assert encoder_out.size(0) == speech.size(0), (
- encoder_out.size(),
- speech.size(0),
- )
- assert encoder_out.size(1) <= encoder_out_lens.max(), (
- encoder_out.size(),
- encoder_out_lens.max(),
- )
+ # 2. Data augmentation
+ if self.specaug is not None and self.training:
+ feats, feats_lengths = self.specaug(feats, feats_lengths)
- if intermediate_outs is not None:
- return (encoder_out, intermediate_outs), encoder_out_lens
+ # 3. Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
+ if self.normalize is not None:
+ feats, feats_lengths = self.normalize(feats, feats_lengths)
- return encoder_out, encoder_out_lens
+ # Pre-encoder, e.g. used for raw input data
+ if self.preencoder is not None:
+ feats, feats_lengths = self.preencoder(feats, feats_lengths)
- def calc_predictor(self, encoder_out, encoder_out_lens):
+ # 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]
- 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, _, pre_peak_index = self.predictor(encoder_out, None, encoder_out_mask, ignore_id=self.ignore_id)
- return pre_acoustic_embeds, pre_token_length
+ # 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
- def cal_decoder_with_predictor(self, encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens):
+ return encoder_out, torch.tensor([encoder_out.size(1)])
- decoder_out, _ = self.decoder(
- encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens
- )
- decoder_out = torch.log_softmax(decoder_out, dim=-1)
- return decoder_out, ys_pad_lens
+ def _calc_att_predictor_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
+ 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, _ = self.predictor(encoder_out,
+ ys_pad,
+ encoder_out_mask,
+ ignore_id=self.ignore_id,
+ mask_chunk_predictor=mask_chunk_predictor,
+ target_label_length=ys_pad_lens,
+ )
+ predictor_alignments, predictor_alignments_len = self.predictor.gen_frame_alignments(pre_alphas,
+ encoder_out_lens)
- def _extract_feats(
- self, speech: torch.Tensor, speech_lengths: torch.Tensor
- ) -> Tuple[torch.Tensor, torch.Tensor]:
- assert speech_lengths.dim() == 1, speech_lengths.shape
+ 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=ys_pad_lens,
+ is_training=self.training,
+ )
+ elif self.encoder.overlap_chunk_cls is not None:
+ encoder_out, encoder_out_lens = self.encoder.overlap_chunk_cls.remove_chunk(encoder_out,
+ encoder_out_lens,
+ chunk_outs=None)
+ # 0. sampler
+ decoder_out_1st = None
+ pre_loss_att = None
+ if self.sampling_ratio > 0.0:
+ if self.step_cur < 2:
+ logging.info("enable sampler in paraformer, sampling_ratio: {}".format(self.sampling_ratio))
+ if self.use_1st_decoder_loss:
+ sematic_embeds, decoder_out_1st, pre_loss_att = \
+ self.sampler_with_grad(encoder_out, encoder_out_lens, ys_pad,
+ ys_pad_lens, pre_acoustic_embeds, scama_mask)
+ else:
+ sematic_embeds, decoder_out_1st = \
+ self.sampler(encoder_out, encoder_out_lens, ys_pad,
+ ys_pad_lens, pre_acoustic_embeds, scama_mask)
+ else:
+ if self.step_cur < 2:
+ logging.info("disable sampler in paraformer, sampling_ratio: {}".format(self.sampling_ratio))
+ sematic_embeds = pre_acoustic_embeds
- # for data-parallel
- speech = speech[:, : speech_lengths.max()]
- if self.frontend is not None:
- # Frontend
- # e.g. STFT and Feature extract
- # data_loader may send time-domain signal in this case
- # speech (Batch, NSamples) -> feats: (Batch, NFrames, Dim)
- feats, feats_lengths = self.frontend(speech, speech_lengths)
- else:
- # No frontend and no feature extract
- feats, feats_lengths = speech, speech_lengths
- return feats, feats_lengths
+ # 1. Forward decoder
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens, scama_mask
+ )
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
- def nll(
- self,
- encoder_out: torch.Tensor,
- encoder_out_lens: torch.Tensor,
- ys_pad: torch.Tensor,
- ys_pad_lens: torch.Tensor,
- ) -> torch.Tensor:
- """Compute negative log likelihood(nll) from transformer-decoder
+ 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)
- Normally, this function is called in batchify_nll.
+ # 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())
- Args:
- encoder_out: (Batch, Length, Dim)
- encoder_out_lens: (Batch,)
- ys_pad: (Batch, Length)
- ys_pad_lens: (Batch,)
- """
- 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
+ return loss_att, acc_att, cer_att, wer_att, loss_pre, pre_loss_att
- # 1. Forward decoder
- decoder_out, _ = self.decoder(
- encoder_out, encoder_out_lens, ys_in_pad, ys_in_lens
- ) # [batch, seqlen, dim]
- batch_size = decoder_out.size(0)
- decoder_num_class = decoder_out.size(2)
- # nll: negative log-likelihood
- nll = torch.nn.functional.cross_entropy(
- decoder_out.view(-1, decoder_num_class),
- ys_out_pad.view(-1),
- ignore_index=self.ignore_id,
- reduction="none",
- )
- nll = nll.view(batch_size, -1)
- nll = nll.sum(dim=1)
- assert nll.size(0) == batch_size
- return nll
+ def sampler(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds, chunk_mask=None):
- def batchify_nll(
- self,
- encoder_out: torch.Tensor,
- encoder_out_lens: torch.Tensor,
- ys_pad: torch.Tensor,
- ys_pad_lens: torch.Tensor,
- batch_size: int = 100,
- ):
- """Compute negative log likelihood(nll) from transformer-decoder
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ ys_pad_masked = ys_pad * tgt_mask[:, :, 0]
+ if self.share_embedding:
+ ys_pad_embed = self.decoder.output_layer.weight[ys_pad_masked]
+ else:
+ ys_pad_embed = self.decoder.embed(ys_pad_masked)
+ with torch.no_grad():
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens, chunk_mask
+ )
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
+ pred_tokens = decoder_out.argmax(-1)
+ nonpad_positions = ys_pad.ne(self.ignore_id)
+ seq_lens = (nonpad_positions).sum(1)
+ same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
+ input_mask = torch.ones_like(nonpad_positions)
+ bsz, seq_len = ys_pad.size()
+ for li in range(bsz):
+ target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
+ if target_num > 0:
+ input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
+ input_mask = input_mask.eq(1)
+ input_mask = input_mask.masked_fill(~nonpad_positions, False)
+ input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
- To avoid OOM, this fuction seperate the input into batches.
- Then call nll for each batch and combine and return results.
- Args:
- encoder_out: (Batch, Length, Dim)
- encoder_out_lens: (Batch,)
- ys_pad: (Batch, Length)
- ys_pad_lens: (Batch,)
- batch_size: int, samples each batch contain when computing nll,
- you may change this to avoid OOM or increase
- GPU memory usage
- """
- total_num = encoder_out.size(0)
- if total_num <= batch_size:
- nll = self.nll(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens)
- else:
- nll = []
- start_idx = 0
- while True:
- end_idx = min(start_idx + batch_size, total_num)
- batch_encoder_out = encoder_out[start_idx:end_idx, :, :]
- batch_encoder_out_lens = encoder_out_lens[start_idx:end_idx]
- batch_ys_pad = ys_pad[start_idx:end_idx, :]
- batch_ys_pad_lens = ys_pad_lens[start_idx:end_idx]
- batch_nll = self.nll(
- batch_encoder_out,
- batch_encoder_out_lens,
- batch_ys_pad,
- batch_ys_pad_lens,
- )
- nll.append(batch_nll)
- start_idx = end_idx
- if start_idx == total_num:
- break
- nll = torch.cat(nll)
- assert nll.size(0) == total_num
- return nll
+ sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
+ input_mask_expand_dim, 0)
+ return sematic_embeds * tgt_mask, decoder_out * tgt_mask
- 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)
+ def sampler_with_grad(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds, chunk_mask=None):
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ ys_pad_masked = ys_pad * tgt_mask[:, :, 0]
+ if self.share_embedding:
+ ys_pad_embed = self.decoder.output_layer.weight[ys_pad_masked]
+ else:
+ ys_pad_embed = self.decoder.embed(ys_pad_masked)
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens, chunk_mask
+ )
+ pre_loss_att = self.criterion_att(decoder_outs[0], ys_pad)
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
+ pred_tokens = decoder_out.argmax(-1)
+ nonpad_positions = ys_pad.ne(self.ignore_id)
+ seq_lens = (nonpad_positions).sum(1)
+ same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
+ input_mask = torch.ones_like(nonpad_positions)
+ bsz, seq_len = ys_pad.size()
+ for li in range(bsz):
+ target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
+ if target_num > 0:
+ input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
+ input_mask = input_mask.eq(1)
+ input_mask = input_mask.masked_fill(~nonpad_positions, False)
+ input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
- # 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
+ sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
+ input_mask_expand_dim, 0)
- # 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]
+ return sematic_embeds * tgt_mask, decoder_out * tgt_mask, pre_loss_att
- 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)
+ def calc_predictor(self, encoder_out, encoder_out_lens):
- # 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())
+ 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,
+ encoder_out_lens+1 if self.predictor.tail_threshold > 0.0 else encoder_out_lens)
- return loss_att, acc_att, cer_att, wer_att, loss_pre
+ 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
- def sampler(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds):
+ return pre_acoustic_embeds, pre_token_length, pre_alphas, pre_peak_index
- tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
- ys_pad *= tgt_mask[:, :, 0]
- ys_pad_embed = self.decoder.embed(ys_pad)
- with torch.no_grad():
- decoder_outs = self.decoder(
- encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens
- )
- decoder_out, _ = decoder_outs[0], decoder_outs[1]
- pred_tokens = decoder_out.argmax(-1)
- nonpad_positions = ys_pad.ne(self.ignore_id)
- seq_lens = (nonpad_positions).sum(1)
- same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
- input_mask = torch.ones_like(nonpad_positions)
- bsz, seq_len = ys_pad.size()
- for li in range(bsz):
- target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
- if target_num > 0:
- input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
- input_mask = input_mask.eq(1)
- input_mask = input_mask.masked_fill(~nonpad_positions, False)
- input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
+ def calc_predictor_chunk(self, encoder_out, cache=None):
- sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
- input_mask_expand_dim, 0)
- return sematic_embeds * tgt_mask, decoder_out * tgt_mask
+ 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 _calc_ctc_loss(
- self,
- encoder_out: torch.Tensor,
- encoder_out_lens: torch.Tensor,
- ys_pad: torch.Tensor,
- ys_pad_lens: torch.Tensor,
- ):
- # Calc CTC loss
- loss_ctc = self.ctc(encoder_out, encoder_out_lens, ys_pad, 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
- # 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
class ParaformerBert(Paraformer):
- """
- Author: Speech Lab, Alibaba Group, China
- Paraformer2: advanced paraformer with LFMMI and bert for non-autoregressive end-to-end speech recognition
- """
+ """
+ Author: Speech Lab of DAMO Academy, Alibaba Group
+ Paraformer2: advanced paraformer with LFMMI and bert for non-autoregressive end-to-end speech recognition
+ """
- def __init__(
- 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,
- joint_network: Optional[torch.nn.Module],
- 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,
- embeds_id: int = 2,
- embeds_loss_weight: float = 0.0,
- embed_dims: int = 768,
- ):
- assert check_argument_types()
- assert 0.0 <= ctc_weight <= 1.0, ctc_weight
- assert 0.0 <= interctc_weight < 1.0, interctc_weight
+ def __init__(
+ self,
+ vocab_size: int,
+ token_list: Union[Tuple[str, ...], List[str]],
+ frontend: Optional[AbsFrontend],
+ specaug: Optional[AbsSpecAug],
+ normalize: Optional[AbsNormalize],
+ encoder: AbsEncoder,
+ 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,
+ embeds_id: int = 2,
+ embeds_loss_weight: float = 0.0,
+ embed_dims: int = 768,
+ preencoder: Optional[AbsPreEncoder] = None,
+ postencoder: Optional[AbsPostEncoder] = None,
+ ):
+ 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,
- joint_network=joint_network,
- 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,
- )
- self.decoder.embeds_id = embeds_id
- decoder_attention_dim = self.decoder.attention_dim
- self.pro_nn = torch.nn.Linear(decoder_attention_dim, embed_dims)
- self.cos = torch.nn.CosineSimilarity(dim=-1, eps=1e-6)
- self.embeds_loss_weight = embeds_loss_weight
- self.length_normalized_loss = length_normalized_loss
+ 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,
+ )
+ self.decoder.embeds_id = embeds_id
+ decoder_attention_dim = self.decoder.attention_dim
+ self.pro_nn = torch.nn.Linear(decoder_attention_dim, embed_dims)
+ self.cos = torch.nn.CosineSimilarity(dim=-1, eps=1e-6)
+ self.embeds_loss_weight = embeds_loss_weight
+ self.length_normalized_loss = length_normalized_loss
- def _calc_embed_loss(self,
- ys_pad: torch.Tensor,
- ys_pad_lens: torch.Tensor,
- embed: torch.Tensor = None,
- embed_lengths: torch.Tensor = None,
- embeds_outputs: torch.Tensor = None,
- ):
- embeds_outputs = self.pro_nn(embeds_outputs)
- tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
- embeds_outputs *= tgt_mask # b x l x d
- embed *= tgt_mask # b x l x d
- cos_loss = 1.0 - self.cos(embeds_outputs, embed)
- cos_loss *= tgt_mask.squeeze(2)
- if self.length_normalized_loss:
- token_num_total = torch.sum(tgt_mask)
- else:
- token_num_total = tgt_mask.size()[0]
- cos_loss_total = torch.sum(cos_loss)
- cos_loss = cos_loss_total / token_num_total
- # print("cos_loss: {}".format(cos_loss))
- return cos_loss
+ def _calc_embed_loss(self,
+ ys_pad: torch.Tensor,
+ ys_pad_lens: torch.Tensor,
+ embed: torch.Tensor = None,
+ embed_lengths: torch.Tensor = None,
+ embeds_outputs: torch.Tensor = None,
+ ):
+ embeds_outputs = self.pro_nn(embeds_outputs)
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ embeds_outputs *= tgt_mask # b x l x d
+ embed *= tgt_mask # b x l x d
+ cos_loss = 1.0 - self.cos(embeds_outputs, embed)
+ cos_loss *= tgt_mask.squeeze(2)
+ if self.length_normalized_loss:
+ token_num_total = torch.sum(tgt_mask)
+ else:
+ token_num_total = tgt_mask.size()[0]
+ cos_loss_total = torch.sum(cos_loss)
+ cos_loss = cos_loss_total / token_num_total
+ # print("cos_loss: {}".format(cos_loss))
+ return cos_loss
+
+ 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]
+ embeds_outputs = None
+ if len(decoder_outs) > 2:
+ embeds_outputs = decoder_outs[2]
+
+ 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, embeds_outputs
+
+ def forward(
+ self,
+ speech: torch.Tensor,
+ speech_lengths: torch.Tensor,
+ text: torch.Tensor,
+ text_lengths: torch.Tensor,
+ embed: torch.Tensor = None,
+ 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, )
+ 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()]
+ if embed is not None:
+ embed = embed[:, :embed_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 = 0.0
+ cos_loss = 0.0
+ 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_ret = self._calc_att_loss(
+ encoder_out, encoder_out_lens, text, text_lengths
+ )
+ loss_att, acc_att, cer_att, wer_att, loss_pre = loss_ret[0], loss_ret[1], loss_ret[2], loss_ret[3], \
+ loss_ret[4]
+ embeds_outputs = None
+ if len(loss_ret) > 5:
+ embeds_outputs = loss_ret[5]
+ if embeds_outputs is not None:
+ cos_loss = self._calc_embed_loss(text, text_lengths, embed, embed_lengths, embeds_outputs)
+
+ # 3. CTC-Att loss definition
+ if self.ctc_weight == 0.0:
+ loss = loss_att + loss_pre * self.predictor_weight + cos_loss * self.embeds_loss_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 + cos_loss * self.embeds_loss_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 > 0.0 else None
+ stats["cos_loss"] = cos_loss.detach().cpu() if cos_loss > 0.0 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 _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)
+class BiCifParaformer(Paraformer):
+ """
+ Paraformer model with an extra cif predictor
+ to conduct accurate timestamp prediction
+ """
- # 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
+ def __init__(
+ self,
+ vocab_size: int,
+ token_list: Union[Tuple[str, ...], List[str]],
+ frontend: Optional[AbsFrontend],
+ specaug: Optional[AbsSpecAug],
+ normalize: Optional[AbsNormalize],
+ encoder: AbsEncoder,
+ 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,
+ preencoder: Optional[AbsPreEncoder] = None,
+ postencoder: Optional[AbsPostEncoder] = None,
+ ):
+ assert 0.0 <= ctc_weight <= 1.0, ctc_weight
+ assert 0.0 <= interctc_weight < 1.0, interctc_weight
- # 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]
- embeds_outputs = None
- if len(decoder_outs) > 2:
- embeds_outputs = decoder_outs[2]
+ 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,
+ )
+ assert isinstance(self.predictor, CifPredictorV3), "BiCifParaformer should use CIFPredictorV3"
- 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)
+ def _calc_pre2_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_token_length2 = self.predictor(encoder_out, ys_pad, encoder_out_mask, ignore_id=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_1st.argmax(dim=-1)
- cer_att, wer_att = self.error_calculator(ys_hat.cpu(), ys_pad.cpu())
+ # loss_pre = self.criterion_pre(ys_pad_lens.type_as(pre_token_length), pre_token_length)
+ loss_pre2 = self.criterion_pre(ys_pad_lens.type_as(pre_token_length2), pre_token_length2)
- return loss_att, acc_att, cer_att, wer_att, loss_pre, embeds_outputs
+ 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)
+ 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)
+ return ds_alphas, ds_cif_peak, us_alphas, us_peaks
+
+ 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
+ )
+
+ loss_pre2 = self._calc_pre2_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 + 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
- def forward(
- self,
- speech: torch.Tensor,
- speech_lengths: torch.Tensor,
- text: torch.Tensor,
- text_lengths: torch.Tensor,
- embed: torch.Tensor = None,
- embed_lengths: torch.Tensor = None,
- ) -> Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]:
- """Frontend + Encoder + Decoder + Calc loss
+class ContextualParaformer(Paraformer):
+ """
+ Paraformer model with contextual hotword
+ """
- 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(), :]
- if embed is not None:
- embed = embed[:, :embed_lengths.max(), :]
+ def __init__(
+ self,
+ vocab_size: int,
+ token_list: Union[Tuple[str, ...], List[str]],
+ frontend: Optional[AbsFrontend],
+ specaug: Optional[AbsSpecAug],
+ normalize: Optional[AbsNormalize],
+ encoder: AbsEncoder,
+ 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,
+ min_hw_length: int = 2,
+ max_hw_length: int = 4,
+ sample_rate: float = 0.6,
+ batch_rate: float = 0.5,
+ double_rate: float = -1.0,
+ target_buffer_length: int = -1,
+ inner_dim: int = 256,
+ bias_encoder_type: str = 'lstm',
+ label_bracket: bool = False,
+ use_decoder_embedding: bool = False,
+ preencoder: Optional[AbsPreEncoder] = None,
+ postencoder: Optional[AbsPostEncoder] = None,
+ ):
+ assert 0.0 <= ctc_weight <= 1.0, ctc_weight
+ assert 0.0 <= interctc_weight < 1.0, interctc_weight
- # 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]
+ 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,
+ )
+ if bias_encoder_type == 'lstm':
+ logging.warning("enable bias encoder sampling and contextual training")
+ self.bias_encoder = torch.nn.LSTM(inner_dim, inner_dim, 1, batch_first=True, dropout=0)
+ self.bias_embed = torch.nn.Embedding(vocab_size, inner_dim)
+ else:
+ logging.error("Unsupport bias encoder type")
- loss_att, acc_att, cer_att, wer_att = None, None, None, None
- loss_ctc, cer_ctc = None, None
- loss_pre = 0.0
- cos_loss = 0.0
- stats = dict()
+ self.min_hw_length = min_hw_length
+ self.max_hw_length = max_hw_length
+ self.sample_rate = sample_rate
+ self.batch_rate = batch_rate
+ self.target_buffer_length = target_buffer_length
+ self.double_rate = double_rate
- # 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
- )
+ if self.target_buffer_length > 0:
+ self.hotword_buffer = None
+ self.length_record = []
+ self.current_buffer_length = 0
+ self.use_decoder_embedding = use_decoder_embedding
- # Collect CTC branch stats
- stats["loss_ctc"] = loss_ctc.detach() if loss_ctc is not None else None
- stats["cer_ctc"] = cer_ctc
+ 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()]
- # 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
+ # 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]
- # 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_att, acc_att, cer_att, wer_att = None, None, None, None
+ loss_ctc, cer_ctc = None, None
+ loss_pre = None
+ stats = dict()
- loss_interctc = loss_interctc / len(intermediate_outs)
+ # 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
+ )
- # calculate whole encoder loss
- loss_ctc = (
- 1 - self.interctc_weight
- ) * loss_ctc + self.interctc_weight * loss_interctc
+ # 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
- # 2b. Attention decoder branch
- if self.ctc_weight != 1.0:
+ # 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_ret = self._calc_att_loss(
- encoder_out, encoder_out_lens, text, text_lengths
- )
- loss_att, acc_att, cer_att, wer_att, loss_pre = loss_ret[0], loss_ret[1], loss_ret[2], loss_ret[3], loss_ret[4]
- embeds_outputs = None
- if len(loss_ret) > 5:
- embeds_outputs = loss_ret[5]
- if embeds_outputs is not None:
- cos_loss = self._calc_embed_loss(text, text_lengths, embed, embed_lengths, embeds_outputs)
+ loss_interctc = loss_interctc / len(intermediate_outs)
- # 3. CTC-Att loss definition
- if self.ctc_weight == 0.0:
- loss = loss_att + loss_pre * self.predictor_weight + cos_loss * self.embeds_loss_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 + cos_loss * self.embeds_loss_weight
+ # calculate whole encoder loss
+ loss_ctc = (
+ 1 - self.interctc_weight
+ ) * loss_ctc + self.interctc_weight * loss_interctc
- # 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 > 0.0 else None
- stats["cos_loss"] = cos_loss.detach().cpu() if cos_loss > 0.0 else None
+ # 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
+ )
- stats["loss"] =torch.clone(loss.detach())
+ # 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
- # 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
+ # 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 _sample_hot_word(self, ys_pad, ys_pad_lens):
+ hw_list = [torch.Tensor([0]).long().to(ys_pad.device)]
+ hw_lengths = [0] # this length is actually for indice, so -1
+ for i, length in enumerate(ys_pad_lens):
+ if length < 2:
+ continue
+ if length > self.min_hw_length + self.max_hw_length + 2 and random.random() < self.double_rate:
+ # sample double hotword
+ _max_hw_length = min(self.max_hw_length, length // 2)
+ # first hotword
+ start1 = random.randint(0, length // 3)
+ end1 = random.randint(start1 + self.min_hw_length - 1, start1 + _max_hw_length - 1)
+ hw_tokens1 = ys_pad[i][start1:end1 + 1]
+ hw_lengths.append(len(hw_tokens1) - 1)
+ hw_list.append(hw_tokens1)
+ # second hotword
+ start2 = random.randint(end1 + 1, length - self.min_hw_length)
+ end2 = random.randint(min(length - 1, start2 + self.min_hw_length - 1),
+ min(length - 1, start2 + self.max_hw_length - 1))
+ hw_tokens2 = ys_pad[i][start2:end2 + 1]
+ hw_lengths.append(len(hw_tokens2) - 1)
+ hw_list.append(hw_tokens2)
+ continue
+ if random.random() < self.sample_rate:
+ if length == 2:
+ hw_tokens = ys_pad[i][:2]
+ hw_lengths.append(1)
+ hw_list.append(hw_tokens)
+ else:
+ start = random.randint(0, length - self.min_hw_length)
+ end = random.randint(min(length - 1, start + self.min_hw_length - 1),
+ min(length - 1, start + self.max_hw_length - 1)) + 1
+ # print(start, end)
+ hw_tokens = ys_pad[i][start:end]
+ hw_lengths.append(len(hw_tokens) - 1)
+ hw_list.append(hw_tokens)
+ # padding
+ hw_list_pad = pad_list(hw_list, 0)
+ 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
+ selected = hw_embed[_ind, hw_lengths]
+ if self.target_buffer_length > 0:
+ _b = selected.shape[0]
+ if self.hotword_buffer is None:
+ self.hotword_buffer = selected
+ self.length_record.append(selected.shape[0])
+ self.current_buffer_length = _b
+ elif self.current_buffer_length + _b < self.target_buffer_length:
+ self.hotword_buffer = torch.cat([self.hotword_buffer.detach(), selected], dim=0)
+ self.current_buffer_length += _b
+ selected = self.hotword_buffer
+ else:
+ self.hotword_buffer = torch.cat([self.hotword_buffer.detach(), selected], dim=0)
+ random_throw = random.randint(self.target_buffer_length // 2, self.target_buffer_length) + 10
+ self.hotword_buffer = self.hotword_buffer[-1 * random_throw:]
+ selected = self.hotword_buffer
+ self.current_buffer_length = selected.shape[0]
+ return selected.squeeze(0).repeat(ys_pad.shape[0], 1, 1).to(ys_pad.device)
+ def sampler(self, encoder_out, encoder_out_lens, ys_pad, ys_pad_lens, pre_acoustic_embeds, contextual_info):
+ tgt_mask = (~make_pad_mask(ys_pad_lens, maxlen=ys_pad_lens.max())[:, :, None]).to(ys_pad.device)
+ ys_pad = ys_pad * tgt_mask[:, :, 0]
+ if self.share_embedding:
+ ys_pad_embed = self.decoder.output_layer.weight[ys_pad]
+ else:
+ ys_pad_embed = self.decoder.embed(ys_pad)
+ with torch.no_grad():
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, pre_acoustic_embeds, ys_pad_lens, contextual_info=contextual_info
+ )
+ decoder_out, _ = decoder_outs[0], decoder_outs[1]
+ pred_tokens = decoder_out.argmax(-1)
+ nonpad_positions = ys_pad.ne(self.ignore_id)
+ seq_lens = (nonpad_positions).sum(1)
+ same_num = ((pred_tokens == ys_pad) & nonpad_positions).sum(1)
+ input_mask = torch.ones_like(nonpad_positions)
+ bsz, seq_len = ys_pad.size()
+ for li in range(bsz):
+ target_num = (((seq_lens[li] - same_num[li].sum()).float()) * self.sampling_ratio).long()
+ if target_num > 0:
+ input_mask[li].scatter_(dim=0, index=torch.randperm(seq_lens[li])[:target_num].cuda(), value=0)
+ input_mask = input_mask.eq(1)
+ input_mask = input_mask.masked_fill(~nonpad_positions, False)
+ input_mask_expand_dim = input_mask.unsqueeze(2).to(pre_acoustic_embeds.device)
+ sematic_embeds = pre_acoustic_embeds.masked_fill(~input_mask_expand_dim, 0) + ys_pad_embed.masked_fill(
+ input_mask_expand_dim, 0)
+ return sematic_embeds * tgt_mask, decoder_out * tgt_mask
+ 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)
+
+ # sample hot word
+ contextual_info = self._sample_hot_word(ys_pad, ys_pad_lens)
+
+ # 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, contextual_info)
+ 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, contextual_info=contextual_info
+ )
+ 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 cal_decoder_with_predictor(self, encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens, hw_list=None, clas_scale=1.0):
+ if hw_list is None:
+ # 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)
+ 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)
+ 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)
+ # hw_embed, _ = torch.nn.utils.rnn.pad_packed_sequence(hw_embed, batch_first=True)
+ contextual_info = h_n.squeeze(0).repeat(encoder_out.shape[0], 1, 1)
+ decoder_outs = self.decoder(
+ encoder_out, encoder_out_lens, sematic_embeds, ys_pad_lens, contextual_info=contextual_info
+ )
+ decoder_out = decoder_outs[0]
+ decoder_out = torch.log_softmax(decoder_out, dim=-1)
+ return decoder_out, ys_pad_lens
+
+ def gen_clas_tf2torch_map_dict(self):
+ tensor_name_prefix_torch = "bias_encoder"
+ tensor_name_prefix_tf = "seq2seq/clas_charrnn"
+
+ tensor_name_prefix_torch_emb = "bias_embed"
+ tensor_name_prefix_tf_emb = "seq2seq"
+
+ map_dict_local = {
+ # in lstm
+ "{}.weight_ih_l0".format(tensor_name_prefix_torch):
+ {"name": "{}/rnn/lstm_cell/kernel".format(tensor_name_prefix_tf),
+ "squeeze": None,
+ "transpose": (1, 0),
+ "slice": (0, 512),
+ "unit_k": 512,
+ }, # (1024, 2048),(2048,512)
+ "{}.weight_hh_l0".format(tensor_name_prefix_torch):
+ {"name": "{}/rnn/lstm_cell/kernel".format(tensor_name_prefix_tf),
+ "squeeze": None,
+ "transpose": (1, 0),
+ "slice": (512, 1024),
+ "unit_k": 512,
+ }, # (1024, 2048),(2048,512)
+ "{}.bias_ih_l0".format(tensor_name_prefix_torch):
+ {"name": "{}/rnn/lstm_cell/bias".format(tensor_name_prefix_tf),
+ "squeeze": None,
+ "transpose": None,
+ "scale": 0.5,
+ "unit_b": 512,
+ }, # (2048,),(2048,)
+ "{}.bias_hh_l0".format(tensor_name_prefix_torch):
+ {"name": "{}/rnn/lstm_cell/bias".format(tensor_name_prefix_tf),
+ "squeeze": None,
+ "transpose": None,
+ "scale": 0.5,
+ "unit_b": 512,
+ }, # (2048,),(2048,)
+
+ # in embed
+ "{}.weight".format(tensor_name_prefix_torch_emb):
+ {"name": "{}/contextual_encoder/w_char_embs".format(tensor_name_prefix_tf_emb),
+ "squeeze": None,
+ "transpose": None,
+ }, # (4235,256),(4235,256)
+ }
+ return map_dict_local
+
+ def clas_convert_tf2torch(self,
+ var_dict_tf,
+ var_dict_torch):
+ map_dict = self.gen_clas_tf2torch_map_dict()
+ var_dict_torch_update = dict()
+ for name in sorted(var_dict_torch.keys(), reverse=False):
+ names = name.split('.')
+ if names[0] == "bias_encoder":
+ name_q = name
+ if name_q in map_dict.keys():
+ name_v = map_dict[name_q]["name"]
+ name_tf = name_v
+ data_tf = var_dict_tf[name_tf]
+ if map_dict[name_q].get("unit_k") is not None:
+ dim = map_dict[name_q]["unit_k"]
+ i = data_tf[:, 0:dim].copy()
+ f = data_tf[:, dim:2 * dim].copy()
+ o = data_tf[:, 2 * dim:3 * dim].copy()
+ g = data_tf[:, 3 * dim:4 * dim].copy()
+ data_tf = np.concatenate([i, o, f, g], axis=1)
+ if map_dict[name_q].get("unit_b") is not None:
+ dim = map_dict[name_q]["unit_b"]
+ i = data_tf[0:dim].copy()
+ f = data_tf[dim:2 * dim].copy()
+ o = data_tf[2 * dim:3 * dim].copy()
+ g = data_tf[3 * dim:4 * dim].copy()
+ data_tf = np.concatenate([i, o, f, g], axis=0)
+ if map_dict[name_q]["squeeze"] is not None:
+ data_tf = np.squeeze(data_tf, axis=map_dict[name_q]["squeeze"])
+ if map_dict[name_q].get("slice") is not None:
+ data_tf = data_tf[map_dict[name_q]["slice"][0]:map_dict[name_q]["slice"][1]]
+ if map_dict[name_q].get("scale") is not None:
+ data_tf = data_tf * map_dict[name_q]["scale"]
+ if map_dict[name_q]["transpose"] is not None:
+ data_tf = np.transpose(data_tf, map_dict[name_q]["transpose"])
+ data_tf = torch.from_numpy(data_tf).type(torch.float32).to("cpu")
+ assert var_dict_torch[name].size() == data_tf.size(), "{}, {}, {} != {}".format(name, name_tf,
+ var_dict_torch[
+ name].size(),
+ data_tf.size())
+ var_dict_torch_update[name] = data_tf
+ logging.info(
+ "torch tensor: {}, {}, loading from tf tensor: {}, {}".format(name, data_tf.size(), name_v,
+ var_dict_tf[name_tf].shape))
+ elif names[0] == "bias_embed":
+ name_tf = map_dict[name]["name"]
+ data_tf = var_dict_tf[name_tf]
+ if map_dict[name]["squeeze"] is not None:
+ data_tf = np.squeeze(data_tf, axis=map_dict[name]["squeeze"])
+ if map_dict[name]["transpose"] is not None:
+ data_tf = np.transpose(data_tf, map_dict[name]["transpose"])
+ data_tf = torch.from_numpy(data_tf).type(torch.float32).to("cpu")
+ assert var_dict_torch[name].size() == data_tf.size(), "{}, {}, {} != {}".format(name, name_tf,
+ var_dict_torch[
+ name].size(),
+ data_tf.size())
+ var_dict_torch_update[name] = data_tf
+ logging.info(
+ "torch tensor: {}, {}, loading from tf tensor: {}, {}".format(name, data_tf.size(), name_tf,
+ var_dict_tf[name_tf].shape))
+
+ return var_dict_torch_update
--
Gitblit v1.9.1