def __init__(self, embedding_matrix, opt):
super(MGANDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(torch.tensor(embedding_matrix, dtype=torch.float))
self.ctx_lstm = DynamicLSTM(opt.embed_dim, opt.hidden_dim, num_layers=1, batch_first=True, bidirectional=True)
self.asp_lstm = DynamicLSTM(opt.embed_dim, opt.hidden_dim, num_layers=1, batch_first=True, bidirectional=True)
self.w_a2c = nn.Parameter(torch.Tensor(2 * opt.hidden_dim, 2 * opt.hidden_dim))
self.w_c2a = nn.Parameter(torch.Tensor(2 * opt.hidden_dim, 2 * opt.hidden_dim))
self.alignment = AlignmentMatrix(opt)
def __init__(self, embedding_matrix, opt):
super(AOADPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(
torch.tensor(embedding_matrix, dtype=torch.float))
self.ctx_lstm = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True,
bidirectional=True)
self.asp_lstm = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True,
bidirectional=True)
def __init__(self, opt, fixed_dim):
super(FixedSALLayer, self).__init__(opt)
self.agg_lstm = DynamicLSTM(fixed_dim,
fixed_dim,
num_layers=1,
batch_first=True)
def __init__(self, opt, times_hidden=1):
super(NormalSALLayer, self).__init__(opt)
self.agg_lstm = DynamicLSTM(times_hidden * opt.hidden_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True)
def __init__(self, embedding_matrix, opt):
super(IANDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(
torch.tensor(embedding_matrix, dtype=torch.float))
self.lstm_context = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True)
self.lstm_aspect = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True)
self.attention_context = Attention(opt.hidden_dim,
score_function='bi_linear')
self.attention_aspect = Attention(opt.hidden_dim,
score_function='bi_linear')
def __init__(self, embedding_matrix, opt):
super(RAMDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(torch.tensor(embedding_matrix, dtype=torch.float))
self.bi_lstm_context = DynamicLSTM(opt.embed_dim, opt.hidden_dim, num_layers=1, batch_first=True,
bidirectional=True)
self.att_linear = nn.Linear(opt.hidden_dim * 2 + 1 + opt.embed_dim * 2, 1)
self.gru_cell = nn.GRUCell(opt.hidden_dim * 2 + 1, opt.hidden_dim)
def __init__(self, embedding_matrix, opt):
super(TNETDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(
torch.tensor(embedding_matrix, dtype=torch.float))
self.lstm1 = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True,
bidirectional=True)
self.lstm2 = DynamicLSTM(opt.embed_dim,
opt.hidden_dim,
num_layers=1,
batch_first=True,
bidirectional=True)
self.convs3 = nn.Conv1d(2 * opt.hidden_dim, 50, 3, padding=1)
self.fc1 = nn.Linear(4 * opt.hidden_dim, 2 * opt.hidden_dim)
def __init__(self, embedding_matrix, opt):
super(BaselineDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(
torch.tensor(embedding_matrix, dtype=torch.float))
self.squeeze_embed = SqueezeEmbedding()
self.slice_lstm = DynamicLSTM(opt.embed_dim * 2,
opt.hidden_dim,
num_layers=1,
batch_first=True)
def __init__(self, embedding_matrix, opt):
super(ATAEDPL, self).__init__(opt)
self.embed = nn.Embedding.from_pretrained(
torch.tensor(embedding_matrix, dtype=torch.float))
self.squeeze_embed = SqueezeEmbedding()
self.lstm = DynamicLSTM(opt.embed_dim * 2,
opt.hidden_dim,
num_layers=1,
batch_first=True)
self.attention = NoQueryAttention(opt.hidden_dim + opt.embed_dim,
score_function='bi_linear')