【PyTorch学习笔记】6.循环神经网络

根据龙良曲Pytorch学习视频整理，视频链接：
【计算机-AI】PyTorch学这个就够了！
(好课推荐)深度学习与PyTorch入门实战——主讲人龙良曲

40.时间序列表示

Sequence representation

• [seq_len, feature_len]
• [word, word_vec]
  one-hot
• [words, word vec]
  sparse
  high-dim
  semantic similarity

word2vec vs Glove

import torch
import torch.nn as nn
from torchnlp.word_to_vector import GloVe

word_to_idx = {'hello': 0, 'world': 1}
lookup_tensor = torch.tensor([word_to_idx['hello']], dtype=torch.long)
embeds = nn.Embedding(2, 5) # 2 words in vocab, 5 dimensional embeddings
hello_embed = embeds(lookup_tensor)
print(hello_embed)
"""
tensor([[ 0.2565, -0.2827, -0.0259, -1.9533,  0.8330]],
       grad_fn=<EmbeddingBackward>)
"""

vectors = GloVe()
print(vectors['hello']) # 2GB文件

torchnlp包安装 pip install pytorch-nlp

41.循环神经网络

Weight sharing
Consistent memory

42.RNN Layer使用

input dim, hidden dim

rnn = nn.RNN(100, 10)   # word_dim, memory/h
print(rnn._parameters.keys())   # odict_keys(['weight_ih_l0', 'weight_hh_l0', 'bias_ih_l0', 'bias_hh_l0'])
print(rnn.weight_hh_l0.shape, rnn.weight_ih_l0.shape)   # torch.Size([10, 10]) torch.Size([10, 100])
print(rnn.bias_hh_l0.shape, rnn.bias_ih_l0.shape)   # torch.Size([10]) torch.Size([10])

42.1 nn.RNN

• .__init__
  (input_size, hidden_size, num_layers)
• out, ht = forward(x, h0)
  x: [seq_len, b, word_vec]
  ho/ht: [num_layers, b, h_dim]
  out: [seq_len, b, h_dim]

Single layer RNN

rnn = nn.RNN(input_size=100, hidden_size=20, num_layers=1)
print(rnn)  # RNN(100, 20)
x = torch.randn(10, 3, 100)
out, h = rnn(x, torch.zeros(1, 3, 20))
print(out.shape, h.shape)   # torch.Size([10, 3, 20]) torch.Size([1, 3, 20])

h是最后一个时间序列所有的memory状态；out是所有时间序列的最后一个memory状态

2 layer RNN

rnn = nn.RNN(100, 10, num_layers=2)   # word_dim, memory/h
print(rnn._parameters.keys())   # odict_keys(['weight_ih_l0', 'weight_hh_l0', 'bias_ih_l0', 'bias_hh_l0', 'weight_ih_l1', 'weight_hh_l1', 'bias_ih_l1', 'bias_hh_l1'])
print(rnn.weight_hh_l0.shape, rnn.weight_ih_l0.shape)   # torch.Size([10, 10]) torch.Size([10, 100])
print(rnn.bias_hh_l0.shape, rnn.bias_ih_l0.shape)   # torch.Size([10]) torch.Size([10])

[T, b, h_dim], [layers, b, h_dim]

rnn = nn.RNN(input_size=100, hidden_size=20, num_layers=4)
print(rnn)  # RNN(100, 20, num_layers=4)
x = torch.randn(10, 3, 100)
out, h = rnn(x)
print(out.shape, h.shape)   # torch.Size([10, 3, 20]) torch.Size([4, 3, 20])

42.2 nn.RNNCell

• __init__
  (input_size, hidden_size, num_layers)
• ht = rnncell(xt, ht_1)
  x: [b, word_vec]
  ht_1/ht: [num_layers, b, h_dim]
  out = torch.stack([h1, h2,…ht])

Functional

cell1 = nn.RNNCell(100, 30)
cell2 = nn.RNNCell(30, 20)
h1 = torch.zeros(3, 30)
h2 = torch.zeros(3, 20)
for xt in x:
    h1 = cell1(xt, h1)
    h2 = cell2(h1, h2)
print(h1.shape)     # torch.Size([3, 30])
print(h2.shape)     # torch.Size([3, 20])

43. 时间序列预测

import numpy as np
import matplotlib.pyplot as plt
import torch
import torch.nn as nn
import torch.optim as optim

num_time_steps = 50
input_size = 1
hidden_size = 16
output_size = 1
lr = 0.01

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()

        self.rnn = nn.RNN(
            input_size=input_size,
            hidden_size=hidden_size,
            num_layers=1,
            batch_first=True
        )
        for p in self.rnn.parameters():
            nn.init.normal_(p, mean=0.0, std=0.001)

        self.linear = nn.Linear(hidden_size, output_size)

    def forward(self, x, hidden_prev):
        out, hidden_prev = self.rnn(x, hidden_prev)
        # [b, seq, h] => [seq, h]   b=1
        out = out.view(-1, hidden_size)
        # [seq, h] => [seq, 1]
        out = self.linear(out)
        # [seq, 1] => [1, seq, 1]
        out = out.unsqueeze(dim=0)
        return out, hidden_prev

model = Net()
criterion = nn.MSELoss()
optimizer = optim.Adam(model.parameters(), lr)

hidden_prev = torch.zeros(1, 1, hidden_size)

for iter in range(6000):
    start = np.random.randint(3, size=1)[0]
    time_steps = np.linspace(start, start + 10, num_time_steps)
    data = np.sin(time_steps)
    data = data.reshape(num_time_steps, 1)
    x = torch.tensor(data[:-1]).float().view(1, num_time_steps - 1, 1)
    y = torch.tensor(data[1:]).float().view(1, num_time_steps - 1, 1)

    output, hidden_prev = model(x, hidden_prev)
    hidden_prev = hidden_prev.detach()  # 改requirse_grad为false，切断反向传播

    loss = criterion(output, y)
    model.zero_grad()
    loss.backward()
    # 检查梯度信息
    # for p in model.parameters():
    #     print(p.grad.norm())
    #     torch.nn.utils.clip_grad_norm_(p, 10)
    optimizer.step()

    if iter % 100 == 0:
        print('Iteration: {} loss {}'.format(iter, loss.item()))

start = np.random.randint(3, size=1)[0]
time_steps = np.linspace(start, start + 10, num_time_steps)
data = np.sin(time_steps)
data = data.reshape(num_time_steps, 1)
x = torch.tensor(data[:-1]).float().view(1, num_time_steps - 1, 1)
y = torch.tensor(data[1:]).float().view(1, num_time_steps - 1, 1)

predictions = []
input = x[:, 0, :]  # [b, seq, feature] => [b, feature]
for _ in range(x.shape[1]):
    input = input.view(1, 1, 1)
    (pred, hidden_prev) = model(input, hidden_prev)
    input = pred    # tensor([[[0.8720]]], grad_fn=<UnsqueezeBackward0>)
    predictions.append(pred.detach().numpy().ravel()[0])

x = x.data.numpy().ravel()
plt.scatter(time_steps[:-1], x.ravel(), s=90)
plt.plot(time_steps[:-1], x.ravel())

plt.scatter(time_steps[1:], predictions)
plt.show()

44.RNN训练难题

• Gradient Exploding
  解决办法：Gradient Clipping
• Gradient Vanishing
  解决办法：LSTM

45.LSTM Layer使用

45.1 nn.LSTM

• __init__
  (input_size, hidden_size, num_layers)

LSTM.forward()

• out, (ht, ct) = lstm(x, [ht_0, ct_0])
  x: [seq_len, b, word_vec]
  h/c: [num_layers, b, h_dim]
  out: [seq_len, b, h_dim]

lstm = nn.LSTM(input_size=100, hidden_size=20, num_layers=4)
print(lstm) # LSTM(100, 20, num_layers=4)
x = torch.rand(10, 3, 100)
out, (h, c) = lstm(x)
print(out.shape, h.shape, c.shape)  # torch.Size([10, 3, 20]) torch.Size([4, 3, 20]) torch.Size([4, 3, 20])

45.2 nn.LSTMCell

• __init__
  (input_size, hidden_size, num_layers)

LSTMCell.forward()

• ht, ct = lstmcell(xt, [ht_0, ct_0])
  xt: [b, word_vec]
  h/c: [ b, h_dim]

Single layer

cell = nn.LSTMCell(input_size=100, hidden_size=20)
print(cell)  # LSTMCell(100, 20)
h = torch.zeros(3, 20)
c = torch.zeros(3, 20)
for xt in x:
    h, c = cell(xt, [h, c])
print(h.shape, c.shape)     # torch.Size([3, 20]) torch.Size([3, 20])

Two layers

cell1 = nn.LSTMCell(input_size=100, hidden_size=30)
cell2 = nn.LSTMCell(input_size=30, hidden_size=20)
h1 = torch.zeros(3, 30)
c1 = torch.zeros(3, 30)
h2 = torch.zeros(3, 20)
c2 = torch.zeros(3, 20)
for xt in x:
    h1, c1 = cell1(xt, [h1, c1])
    h2, c2 = cell2(h1, [h2, c2])
print(h2.shape, c2.shape)   # torch.Size([3, 20]) torch.Size([3, 20])

46.情感分类实战

Google CoLab

• Continuous 12 hours
• free K80 for GPU
• no need to cross GFW

没有谷歌账号不能使用Colaboratory，码一下

import torch
import torch.nn as nn
import numpy as np
from torchtext import data, datasets

# load dataset
TEXT = data.Field(tokenize='spacy')
LABEL = data.LabelField(dtype=torch.float)
train_data, test_data = datasets.IMDB.split(TEXT, LABEL)

print('len of train data:', len(train_data))
print('len of test data:', len(test_data))
print(train_data.examples[15].text)
print(train_data.examples[15].label)

class RNN(nn.Module):
    def __init__(self, vocab_size, embedding_dim, hidden_dim):
        super(RNN, self).__init__()

        # [0-10001] => [100]
        self.embedding = nn.Embedding(vocab_size, embedding_dim)
        # [100] => [256]
        self.rnn = nn.LSTM(embedding_dim, hidden_dim, num_layers=2, bidirectional=True, dropout=0.5)
        # [25*2] => [1]
        self.fc = nn.Linear(hidden_dim*2, 1)
        self.dropout = nn.Dropout(0.5)

    def forward(self, x):
        # [seq, b, 1] => [seq, b, 100]
        embedding = self.dropout(self.embedding(x))
        # output: [seq, b, hid_dim*2]
        # hidden/h, [num_layer*2, b, hid_dim]
        # cell/c: [num_layer*2, b, hid_dim]
        output, (hidden, cell) = self.rnn(embedding)
        # [num_layer*2, b, hid_dim] => 2 of [b, hid_dim] => [b, hid_dim*2]
        hidden = torch.cat([hidden[-2], hidden[-1]], dim=1)
        # [b, hid_dim*2] => [b, 1]
        hidden = self.dropout(hidden)
        out = self.fc(hidden)
        return out

# load word embedding
rnn = RNN(len(TEXT.vocab), 100, 256)
pretrained_embedding = TEXT.vocab.vectors
print('pretrained_embedding:', pretrained_embedding.shape)
rnn.embedding.weight.data.copy_(pretrained_embedding)
print('embedding layer inited.')


def binary_acc(preds, y):
    preds = torch.round((torch.sigmoid(preds)))
    correct = torch.eq(preds, y).float()
    acc = correct.sum() / len(correct)
    return acc

def eval(rnn, iterator, criteon):
    avg_acc = []
    rnn.eval()
    with torch.no_grad():
        for batch in iterator:
            # [b, 1] => [b]
            pred = rnn(batch.text).squeeze(1)
            loss = criteon(pred, batch.label)
            acc = binary_acc(pred, batch.label).item()
            avg_acc.append(acc)
    avg_acc = np.array(avg_acc).mean()
    print('>>>test:', avg_acc)

def train(rnn, iterator, optimizer, criteon):
    avg_acc = []
    rnn.train()

    for i, batch in enumerate(iterator):
        # [seq, b] => [b, 1] => [b]
        pred = rnn(batch.text).squeeze(1)
        loss = criteon(pred, batch.label)
        acc = binary_acc(pred, batch.label).item()
        avg_acc.append(acc)

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

def main():
    pass

if __name__ == '__main__':
    main()

只是抄了一下代码，并没有动脑子