Source code for openspeech.decoders.transformer_transducer_decoder

# MIT License
#
# Copyright (c) 2021 Soohwan Kim and Sangchun Ha and Soyoung Cho
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

import torch
import torch.nn as nn
import numpy as np
from typing import Tuple

from openspeech.decoders import OpenspeechDecoder
from openspeech.encoders.transformer_transducer_encoder import TransformerTransducerEncoderLayer
from openspeech.modules import (
    PositionalEncoding,
    get_attn_pad_mask,
    get_attn_subsequent_mask,
)


[docs]class TransformerTransducerDecoder(OpenspeechDecoder): r""" Converts the label to higher feature values Args: num_classes (int): the number of vocabulary model_dim (int): the number of features in the label encoder (default : 512) d_ff (int): the number of features in the feed forward layers (default : 2048) num_layers (int): the number of label encoder layers (default: 2) num_heads (int): the number of heads in the multi-head attention (default: 8) dropout (float): dropout probability of label encoder (default: 0.1) max_positional_length (int): Maximum length to use for positional encoding (default : 5000) pad_id (int): index of padding (default: 0) sos_id (int): index of the start of sentence (default: 1) eos_id (int): index of the end of sentence (default: 2) Inputs: inputs, inputs_lens - **inputs**: Ground truth of batch size number - **inputs_lens**: Tensor of target lengths Returns: (torch.FloatTensor, torch.FloatTensor) * outputs (torch.FloatTensor): ``(batch, seq_length, dimension)`` * input_lengths (torch.FloatTensor): ``(batch)`` Reference: Qian Zhang et al.: Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss https://arxiv.org/abs/2002.02562 """ def __init__( self, num_classes: int, model_dim: int = 512, d_ff: int = 2048, num_layers: int = 2, num_heads: int = 8, dropout: float = 0.1, max_positional_length: int = 5000, pad_id: int = 0, sos_id: int = 1, eos_id: int = 2, ) -> None: super(TransformerTransducerDecoder, self).__init__() self.embedding = nn.Embedding(num_classes, model_dim) self.scale = np.sqrt(model_dim) self.positional_encoding = PositionalEncoding(model_dim, max_positional_length) self.input_dropout = nn.Dropout(p=dropout) self.pad_id = pad_id self.sos_id = sos_id self.eos_id = eos_id self.encoder_layers = nn.ModuleList([ TransformerTransducerEncoderLayer( model_dim, d_ff, num_heads, dropout ) for _ in range(num_layers) ])
[docs] def forward( self, inputs: torch.Tensor, input_lengths: torch.Tensor, hidden_states: torch.Tensor = None, ) -> Tuple[torch.Tensor, torch.Tensor]: r""" Forward propagate a `inputs` for label encoder. Args: inputs (torch.LongTensor): A input sequence passed to label encoder. Typically inputs will be a padded `LongTensor` of size ``(batch, target_length)`` input_lengths (torch.LongTensor): The length of input tensor. ``(batch)`` Returns: * outputs (Tensor): ``(batch, seq_length, dimension)`` * input_lengths (Tensor): ``(batch)`` """ self_attn_mask = None batch = inputs.size(0) inputs = inputs[inputs != self.eos_id].view(batch, -1) if len(inputs.size()) == 1: # validate, evaluation inputs = inputs.unsqueeze(1) target_lens = inputs.size(1) embedding_output = self.embedding(inputs) * self.scale positional_encoding_output = self.positional_encoding(target_lens) inputs = embedding_output + positional_encoding_output else: # train inputs = inputs[inputs != self.eos_id].view(batch, -1) target_lengths = inputs.size(1) embedding_output = self.embedding(inputs) * self.scale positional_encoding_output = self.positional_encoding(target_lengths) inputs = embedding_output + positional_encoding_output dec_self_attn_pad_mask = get_attn_pad_mask(inputs, input_lengths, inputs.size(1)) dec_self_attn_subsequent_mask = get_attn_subsequent_mask(inputs) self_attn_mask = torch.gt((dec_self_attn_pad_mask + dec_self_attn_subsequent_mask), 0) outputs = self.input_dropout(inputs) for encoder_layer in self.encoder_layers: outputs, _ = encoder_layer(outputs, self_attn_mask) return outputs, input_lengths