import torch import torch.nn as nn import torch.nn.functional as F import numpy as np def gather_last_relevant_hidden(hiddens, seq_lens): """Extract and collect the last relevant hidden state based on the sequence length.""" seq_lens = seq_lens.long().detach().cpu().numpy() - 1 out = [] for batch_index, column_index in enumerate(seq_lens): out.append(hiddens[batch_index, column_index]) return torch.stack(out) class SkimlitModel(nn.Module): def __init__(self, embedding_dim, vocab_size, hidden_dim, n_layers, linear_output, num_classes, pretrained_embeddings=None, padding_idx=0): super(SkimlitModel, self).__init__() # Initalizing embeddings if pretrained_embeddings is None: self.embeddings = nn.Embedding(num_embeddings=vocab_size, embedding_dim=embedding_dim) else: pretrained_embeddings = torch.from_numpy(pretrained_embeddings).float() self.embeddings = nn.Embedding(num_embeddings=vocab_size, embedding_dim=embedding_dim, _weight=pretrained_embeddings, padding_idx=padding_idx) # LSTM layers self.lstm1 = nn.LSTM(embedding_dim, hidden_dim, num_layers=n_layers, batch_first=True, bidirectional=True) # FC layers self.fc_text = nn.Linear(2*hidden_dim, linear_output) self.fc_line_num = nn.Linear(20, 64) self.fc_total_line = nn.Linear(24, 64) self.fc_final = nn.Linear((64+64+linear_output), num_classes) self.dropout = nn.Dropout(0.3) def forward(self, inputs): x_in, seq_lens, line_nums, total_lines = inputs x_in = self.embeddings(x_in) # RNN outputs out, b_n = self.lstm1(x_in) x_1 = gather_last_relevant_hidden(hiddens=out, seq_lens=seq_lens) # FC layers output x_1 = F.relu(self.fc_text(x_1)) x_2 = F.relu(self.fc_line_num(line_nums)) x_3 = F.relu(self.fc_total_line(total_lines)) x = torch.cat((x_1, x_2, x_3), dim=1) x = self.dropout(x) x = self.fc_final(x) return x