def __init__(self,
rnn_type,
ntoken,
ninp,
nhid,
nlayers,
dropout=0.5,
tie_weights=False):
super(RNNModel, self).__init__()
self.drop = nn.Dropout(dropout)
self.encoder = nn.Embedding(ntoken, ninp)
if rnn_type in ['LSTM', 'GRU']:
self.rnn = getattr(nn, rnn_type)(ninp,
nhid,
nlayers,
dropout=dropout)
elif rnn_type == 'DRNN':
self.rnn = drnn.DRNN(ninp, nhid, nlayers, 0, 'GRU')
self.decoder = nn.Linear(nhid, ntoken)
if tie_weights:
if nhid != ninp:
raise ValueError(
'When using the tied flag, nhid must be equal to emsize')
self.decoder.weight = self.encoder.weight
try:
self.init_weights()
except AttributeError:
pass
self.rnn_type = rnn_type
self.nhid = nhid
self.nlayers = nlayers
def __init__(self):
torch.nn.Module.__init__(self)
self.embed = torch.nn.Embedding(10, 10)
self.drnn = drnn.DRNN(10, 10, 9, 'RNN')
self.project = torch.nn.Linear(10, 8)
def test(self):
model = drnn.DRNN(10, 10, 4, 0, 'GRU')
hidden = []
for i in range(4):
hidden.append(torch.autograd.Variable(torch.randn(2**i, 3, 10)))
x = torch.autograd.Variable(torch.randn(24, 3, 10))
hidden = model(x, hidden)
def test(self):
model = drnn.DRNN(10, 10, 4, 0, 'GRU')
x = torch.randn(23, 3, 10)
out = model(x)[0]
self.assertTrue(out.size(0) == 23)
self.assertTrue(out.size(1) == 3)
self.assertTrue(out.size(2) == 10)
def test(self):
model = drnn.DRNN(10, 10, 4, 0, 'GRU')
x = torch.autograd.Variable(torch.randn(23, 3, 10))
hidden = model(x)[1]
self.assertEqual(len(hidden), 4)
for hid in hidden:
print(hid.size())
def __init__(self,
n_inputs,
n_hidden,
n_layers,
n_classes,
cell_type="GRU"):
super(Classifier, self).__init__()
self.drnn = drnn.DRNN(n_inputs,
n_hidden,
n_layers,
cell_type=cell_type)
self.linear = nn.Linear(n_hidden, n_classes)
def test(self):
model = drnn.DRNN(10, 10, 4, 0, 'GRU')
x = torch.autograd.Variable(torch.randn(24, 3, 10))
split_x = model._prepare_inputs(x, 2)
second_block = x[1::2]
check = split_x[:, x.size(1):, :]
self.assertTrue((second_block == check).all())
unsplit_x = model._split_outputs(split_x, 2)
self.assertTrue((x == unsplit_x).all())
