import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
class CNN(nn.Module):
def __init__(self, emb_dim, emb_num, config):
super(CNN, self).__init__()
if config.emb_mat is None:
V = emb_num
D = emb_dim
else:
V,D = config.emb_mat.shape
C = config.cls_num
Ci = 1
Co = config.kernel_num
Ks = config.kernel_sizes
self.static = config.static_emb
self.embed = nn.Embedding(V, D, padding_idx=config.pad_wid)
# self.convs1 = [nn.Conv2d(Ci, Co, (K, D)) for K in Ks]
self.convs1 = nn.ModuleList([nn.Conv2d(Ci, Co, (K, D)) for K in Ks])
'''
self.conv13 = nn.Conv2d(Ci, Co, (3, D))
self.conv14 = nn.Conv2d(Ci, Co, (4, D))
self.conv15 = nn.Conv2d(Ci, Co, (5, D))
'''
self.dropout = nn.Dropout(config.dropout)
self.fc1 = nn.Linear(len(Ks) * Co, C)
def conv_and_pool(self, x, conv):
x = F.relu(conv(x)).squeeze(3) # (N, Co, W)
x = F.max_pool1d(x, x.size(2)).squeeze(2)
return x
def forward(self, x):
x = self.embed(x) # (N, W, D)
if self.static:
x = Variable(x)
x = x.unsqueeze(1) # (N, Ci, W, D)
x = [F.relu(conv(x)).squeeze(3) for conv in self.convs1] # [(N, Co, W), ...]*len(Ks)
x = [F.max_pool1d(i, i.size(2)).squeeze(2) for i in x] # [(N, Co), ...]*len(Ks)
x = torch.cat(x, 1)
'''
x1 = self.conv_and_pool(x,self.conv13) #(N,Co)
x2 = self.conv_and_pool(x,self.conv14) #(N,Co)
x3 = self.conv_and_pool(x,self.conv15) #(N,Co)
x = torch.cat((x1, x2, x3), 1) # (N,len(Ks)*Co)
'''
x = self.dropout(x) # (N, len(Ks)*Co)
logit = self.fc1(x) # (N, C)
return logit
class MultiFC(nn.Module):
"""
Applies fully connected layers to an input vector
"""
def __init__(self, input_size, hidden_size, output_size, num_hidden_layers=0, short_cut=False):
super(MultiFC, self).__init__()
self.num_hidden_layers = num_hidden_layers
self.short_cut = short_cut
if num_hidden_layers == 0:
self.fc = nn.Linear(input_size, output_size)
else:
self.fc_layers = nn.ModuleList()
self.fc_input = nn.Linear(input_size, hidden_size)
self.fc_output = nn.Linear(hidden_size, output_size)
if short_cut:
self.es = nn.Linear(input_size, output_size)
for i in range(1, self.num_hidden_layers):
self.fc_layers.append(nn.Linear(hidden_size, hidden_size))
def forward(self, input_var):
if self.num_hidden_layers == 0:
out = self.fc(input_var)
else:
x = F.tanh(self.fc_input(input_var))
for i in range(1, self.num_hidden_layers):
x = F.tanh(self.fc_layers[i](x))
out = self.fc_output(x)
if self.short_cut:
out = out + self.es(input_var)
return out
