PyTorch Lecture12：RNN1 - Basics (代码12_3_hello_rnn_seq)

# 12_3_hello_rnn_seq

import torch
import torch.nn as nn
from torch.autograd import Variable

torch.manual_seed(777)  # reproducibility
#            0    1    2    3    4
idx2cahr = ['h', 'i', 'e', 'l', 'o']

# Tech 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

# As we have one batch of samples, we will change them to variables only once
inputs = Variable(torch.Tensor(x_one_hot))
labels = Variable(torch.LongTensor(y_data))

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|
num_layers = 1  # one -layer rnn


class RNN(nn.Module):
    def __init__(self, num_calsses, 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

        self.rnn = nn.RNN(input_size=5, hidden_size=5, batch_first=True)

    def forward(self, x):
        # Initialize hidden and cell states
        # (num_layers*num_directions,batch,hidden_state) for batch_first =True
        h_0 = Variable(torch.zeros(self.num_layers, x.size(0), self.hidden_size))
        # Propagate input through RNN
        # Input:(batch,seq_len,input_size)
        out, _ = self.rnn(x, h_0)
        return out.view(-1, num_classes)


# Instantitate RNN model
rnn = RNN(num_classes, input_size, hidden_size, num_layers)
print(rnn)

# Set loss and optimizer function
# CrossEntropyloss =LogSoftmax + NLLLoss

criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(rnn.parameters(), lr=0.1)

# Trian the model
for epoch in range(10):
    outputs = rnn(inputs)
    optimizer.zero_grad()
    loss = criterion(outputs, labels)
    loss.backward()
    optimizer.step()
    _, idx = outputs.max(1)
    idx = idx.data.numpy()
    result_str = [idx2cahr[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")