RNN循环神经网络（进阶）
1.GRU（门控）模型
1.1.概念
1.1.1.RNN存在的问题：梯度较容易出现衰减或爆炸（BPTT）
1.1.2.GRU⻔控循环神经⽹络：捕捉时间序列中时间步距离较⼤的依赖关系
重置⻔有助于捕捉时间序列⾥短期的依赖关系；
更新⻔有助于捕捉时间序列⾥⻓期的依赖关系。
1.2.GRU模型从零实现
1.2.1.载入数据集
import numpy as np
import torch
from torch import nn, optim
import torch.nn.functional as F
import sys
sys.path.append("../input/")
import d2l_jay9460 as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
(corpus_indices, char_to_idx, idx_to_char, vocab_size) = d2l.load_data_jay_lyrics()
1.2.2.初始化参数
num_inputs, num_hiddens, num_outputs = vocab_size, 256, vocab_size
print('will use', device)
def get_params():
def _one(shape):
ts = torch.tensor(np.random.normal(0, 0.01, size=shape), device=device, dtype=torch.float32) #正态分布
return torch.nn.Parameter(ts, requires_grad=True)
def _three():
return (_one((num_inputs, num_hiddens)),
_one((num_hiddens, num_hiddens)),
torch.nn.Parameter(torch.zeros(num_hiddens, device=device, dtype=torch.float32), requires_grad=True))
W_xz, W_hz, b_z = _three() # 更新门参数
W_xr, W_hr, b_r = _three() # 重置门参数
W_xh, W_hh, b_h = _three() # 候选隐藏状态参数
# 输出层参数
W_hq = _one((num_hiddens, num_outputs))
b_q = torch.nn.Parameter(torch.zeros(num_outputs, device=device, dtype=torch.float32), requires_grad=True)
return nn.ParameterList([W_xz, W_hz, b_z, W_xr, W_hr, b_r, W_xh, W_hh, b_h, W_hq, b_q])
def init_gru_state(batch_size, num_hiddens, device): #隐藏状态初始化
return (torch.zeros((batch_size, num_hiddens), device=device), )
1.2.3.GRU模型
def gru(inputs, state, params):
W_xz, W_hz, b_z, W_xr, W_hr, b_r, W_xh, W_hh, b_h, W_hq, b_q = params
H, = state
outputs = []
for X in inputs:
Z = torch.sigmoid(torch.matmul(X, W_xz) + torch.matmul(H, W_hz) + b_z)
R = torch.sigmoid(torch.matmul(X, W_xr) + torch.matmul(H, W_hr) + b_r)
H_tilda = torch.tanh(torch.matmul(X, W_xh) + R * torch.matmul(H, W_hh) + b_h)
H = Z * H + (1 - Z) * H_tilda
Y = torch.matmul(H, W_hq) + b_q
outputs.append(Y)
return outputs, (H,)
1.2.4.训练和测试模型
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
d2l.train_and_predict_rnn(gru, get_params, init_gru_state, num_hiddens,
vocab_size, device, corpus_indices, idx_to_char,
char_to_idx, False, num_epochs, num_steps, lr,
clipping_theta, batch_size, pred_period, pred_len,
prefixes)
1.3.GRU模型实现PyTorch简洁实现
num_hiddens=256
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
lr = 1e-2 # 注意调整学习率
gru_layer = nn.GRU(input_size=vocab_size, hidden_size=num_hiddens)
model = d2l.RNNModel(gru_layer, vocab_size).to(device)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
corpus_indices, idx_to_char, char_to_idx,
num_epochs, num_steps, lr, clipping_theta,
batch_size, pred_period, pred_len, prefixes)
2.LSTM（长短期记忆）模型
2.1.概念
2.1.1.长短期记忆long short-term memory :
遗忘门:控制上一时间步的记忆细胞 输入门:控制当前时间步的输入
输出门:控制从记忆细胞到隐藏状态
记忆细胞：⼀种特殊的隐藏状态的信息的流动
2.1.2. 结构图
2.2.LSTM模型从零实现
2.2.1.初始化参数
num_inputs, num_hiddens, num_outputs = vocab_size, 256, vocab_size
print('will use', device)
def get_params():
def _one(shape):
ts = torch.tensor(np.random.normal(0, 0.01, size=shape), device=device, dtype=torch.float32)
return torch.nn.Parameter(ts, requires_grad=True)
def _three():
return (_one((num_inputs, num_hiddens)),
_one((num_hiddens, num_hiddens)),
torch.nn.Parameter(torch.zeros(num_hiddens, device=device, dtype=torch.float32), requires_grad=True))
W_xi, W_hi, b_i = _three() # 输入门参数
W_xf, W_hf, b_f = _three() # 遗忘门参数
W_xo, W_ho, b_o = _three() # 输出门参数
W_xc, W_hc, b_c = _three() # 候选记忆细胞参数
# 输出层参数
W_hq = _one((num_hiddens, num_outputs))
b_q = torch.nn.Parameter(torch.zeros(num_outputs, device=device, dtype=torch.float32), requires_grad=True)
return nn.ParameterList([W_xi, W_hi, b_i, W_xf, W_hf, b_f, W_xo, W_ho, b_o, W_xc, W_hc, b_c, W_hq, b_q])
def init_lstm_state(batch_size, num_hiddens, device):
return (torch.zeros((batch_size, num_hiddens), device=device),
torch.zeros((batch_size, num_hiddens), device=device))
2.2.2.LSTM模型
def lstm(inputs, state, params):
[W_xi, W_hi, b_i, W_xf, W_hf, b_f, W_xo, W_ho, b_o, W_xc, W_hc, b_c, W_hq, b_q] = params
(H, C) = state
outputs = []
for X in inputs:
I = torch.sigmoid(torch.matmul(X, W_xi) + torch.matmul(H, W_hi) + b_i)
F = torch.sigmoid(torch.matmul(X, W_xf) + torch.matmul(H, W_hf) + b_f)
O = torch.sigmoid(torch.matmul(X, W_xo) + torch.matmul(H, W_ho) + b_o)
C_tilda = torch.tanh(torch.matmul(X, W_xc) + torch.matmul(H, W_hc) + b_c)
C = F * C + I * C_tilda
H = O * C.tanh()
Y = torch.matmul(H, W_hq) + b_q
outputs.append(Y)
return outputs, (H, C)
2.2.3.训练模型
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
d2l.train_and_predict_rnn(lstm, get_params, init_lstm_state, num_hiddens,
vocab_size, device, corpus_indices, idx_to_char,
char_to_idx, False, num_epochs, num_steps, lr,
clipping_theta, batch_size, pred_period, pred_len,
prefixes)
2.3.LSTM模型PyTorch简洁实现
num_hiddens=256
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
lr = 1e-2 # 注意调整学习率
lstm_layer = nn.LSTM(input_size=vocab_size, hidden_size=num_hiddens)
model = d2l.RNNModel(lstm_layer, vocab_size)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
corpus_indices, idx_to_char, char_to_idx,
num_epochs, num_steps, lr, clipping_theta,
batch_size, pred_period, pred_len, prefixes)
3.深度循环神经网络
3.1.
num_hiddens=256
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
lr = 1e-2 # 注意调整学习率
gru_layer = nn.LSTM(input_size=vocab_size, hidden_size=num_hiddens,num_layers=2)
model = d2l.RNNModel(gru_layer, vocab_size).to(device)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
corpus_indices, idx_to_char, char_to_idx,
num_epochs, num_steps, lr, clipping_theta,
batch_size, pred_period, pred_len, prefixes)
gru_layer = nn.LSTM(input_size=vocab_size, hidden_size=num_hiddens,num_layers=6)
model = d2l.RNNModel(gru_layer, vocab_size).to(device)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
corpus_indices, idx_to_char, char_to_idx,
num_epochs, num_steps, lr, clipping_theta,
batch_size, pred_period, pred_len, prefixes)
3.2.前一层（l-1层）的输出O_l-1作为下一层（l层）的输入X_l；深层RNN可以提取更高层的信息，但是层数越多，模型越复杂，对数据的要求也会更高。
4.双向循环神经网络
4.1.
num_hiddens=128
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e-2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']
lr = 1e-2 # 注意调整学习率
gru_layer = nn.GRU(input_size=vocab_size, hidden_size=num_hiddens,bidirectional=True)
model = d2l.RNNModel(gru_layer, vocab_size).to(device)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
corpus_indices, idx_to_char, char_to_idx,
num_epochs, num_steps, lr, clipping_theta,
batch_size, pred_period, pred_len, prefixes)