def __init__(self, vocab, pos, rels, enum_word, options, onto, cpos):
super(HybridModel, self).__init__()
random.seed(2)
dims = options.wembedding_dims + options.pembedding_dims + options.cembedding_dims + options.oembedding_dims
self.shared = nn.LSTM(dims,
options.lstm_dims,
batch_first=True,
bidirectional=True)
model0 = GraphModel(vocab, pos, rels, enum_word, options, onto, cpos,
self.shared)
model1 = TransitionModel(vocab, pos, rels, enum_word, options, onto,
cpos, self.shared)
self.graphModel = model0.cuda() if torch.cuda.is_available(
) else model0
self.transitionModel = model1.cuda() if torch.cuda.is_available(
) else model1
self.graphTrainer = get_optim(options.optim,
self.graphModel.parameters())
self.transitionTrainer = get_optim(options.optim,
self.transitionModel.parameters())
classifier = LinearClassifier(options.lstm_dims * 4)
self.classifier = classifier.cuda() if torch.cuda.is_available(
) else classifier
self.loss = nn.CrossEntropyLoss()
self.optimizer = optim.Adam(classifier.parameters(), lr=0.01)
def __init__(self, keyword, src_id, dst_id, in_dim, hid_c, src_len,
n_layers, device):
super(PredictModel, self).__init__()
self.oneHotEmbed = OneHotProcess(in_dim, hid_c)
if keyword == "Graph":
self.model = GraphModel(src_id, dst_id, src_len * hid_c,
src_len * hid_c, device)
elif keyword == "SG":
self.model = SGModel(src_id, dst_id, src_len * hid_c,
src_len * hid_c, n_layers, device)
elif keyword == "SAGE":
self.model = SAGEModel(src_id, dst_id, src_len * hid_c,
src_len * hid_c, n_layers, device)
else:
raise KeyError("Keyword is not defined! ")
self.linear = nn.Linear(src_len * hid_c, in_dim)
def _getGraph():
loader = getLoaderPrimaryForm()
g = GraphModel(loader, degree)
g.processGraphs(slice(51555))
return g
