In [13]:
# Lab 12 RNN
import torch
import torch.nn as nn
from torch.autograd import Variable
torch.manual_seed(777) # reproducibility
# hyperparameters
learning_rate = 0.1
num_epochs = 15
idx2char = ['h', 'i', 'e', 'l', 'o']
# Teach hello: hihell -> ihello
x_data = [[0, 1, 0, 2, 3, 3]] # hihell
x_one_hot = [[[1, 0, 0, 0, 0], # h 0
[0, 1, 0, 0, 0], # i 1
[1, 0, 0, 0, 0], # h 0
[0, 0, 1, 0, 0], # e 2
[0, 0, 0, 1, 0], # l 3
[0, 0, 0, 1, 0]]] # l 3
y_data = [1, 0, 2, 3, 3, 4] # ihello
In [14]:
# As we have one batch of samples, we will change them to variables only once
inputs = torch.Tensor(x_one_hot)
labels = torch.LongTensor(y_data)
inputs = Variable(inputs)
labels = Variable(labels)
num_classes = 5
input_size = 5 # one-hot size
hidden_size = 5 # output from the LSTM. 5 to directly predict one-hot
batch_size = 1 # one sentence
sequence_length = 6 # |ihello| == 6
num_layers = 1 # one-layer rnn
In [15]:
labels
Out[15]:
In [16]:
class RNN(nn.Module):
def __init__(self, num_classes, input_size, hidden_size, num_layers):
super(RNN, self).__init__()
self.num_classes = num_classes
self.num_layers = num_layers
self.input_size = input_size
self.hidden_size = hidden_size
self.sequence_length = sequence_length
# Set parameters for RNN block
# Note: batch_first=False by default.
# When true, inputs are (batch_size, sequence_length, input_dimension)
# instead of (sequence_length, batch_size, input_dimension)
self.rnn = nn.RNN(input_size=input_size, hidden_size=hidden_size,
num_layers=num_layers, batch_first=True)
# Fully connected layer to obtain outputs corresponding to the number
# of classes
self.fc = nn.Linear(hidden_size, num_classes)
def forward(self, x):
# Initialize hidden and cell states
h_0 = Variable(torch.zeros(
x.size(0), self.num_layers, self.hidden_size))
# Reshape input
x.view(x.size(0), self.sequence_length, self.input_size)
# Propagate input through RNN
# Input: (batch, seq_len, input_size)
# h_0: (batch, num_layers * num_directions, hidden_size)
out, _ = self.rnn(x, h_0)
# Reshape output from (batch, seq_len, hidden_size) to (batch *
# seq_len, hidden_size)
out = out.view(-1, self.hidden_size)
# Return outputs applied to fully connected layer
out = self.fc(out)
return out
In [17]:
# Instantiate RNN model
rnn = RNN(num_classes, input_size, hidden_size, num_layers)
In [18]:
# Set loss and optimizer function
criterion = torch.nn.CrossEntropyLoss() # Softmax is internally computed.
optimizer = torch.optim.Adam(rnn.parameters(), lr=learning_rate)
In [19]:
# Train the model
for epoch in range(num_epochs):
outputs = rnn(inputs)
optimizer.zero_grad()
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
_, idx = outputs.max(1)
idx = idx.data.numpy()
result_str = [idx2char[c] for c in idx.squeeze()]
print("epoch: %d, loss: %1.3f" % (epoch + 1, loss.data[0]))
print("Predicted string: ", ''.join(result_str))
print("Learning finished!")
In [ ]:
In [ ]: