目录
6.3 LSTM的记忆能力实验
6.3.1 模型构建
6.3.1.1 LSTM层
6.3.1.2 模型汇总
6.3.2 模型训练
6.3.2.1 训练指定长度的数字预测模型
6.3.2.2 多组训练
6.3.2.3 损失曲线展示
【思考题1】LSTM与SRN实验结果对比,谈谈看法。(选做)
6.3.3 模型评价
6.3.3.1 在测试集上进行模型评价
6.3.3.2 模型在不同长度的数据集上的准确率变化图
【思考题2】LSTM与SRN在不同长度数据集上的准确度对比,谈谈看法。(选做)
6.3.3.3 LSTM模型门状态和单元状态的变化
【思考题3】分析LSTM中单元状态和门数值的变化图,并用自己的话解释该图。
全面总结RNN(必做)
长短期记忆网络(Long Short-Term Memory Network,LSTM)是一种可以有效缓解长程依赖问题的循环神经网络.LSTM 的特点是引入了一个新的内部状态(Internal State)和门控机制(Gating Mechanism).不同时刻的内部状态以近似线性的方式进行传递,从而缓解梯度消失或梯度爆炸问题.同时门控机制进行信息筛选,可以有效地增加记忆能力.例如,输入门可以让网络忽略无关紧要的输入信息,遗忘门可以使得网络保留有用的历史信息.在上一节的数字求和任务中,如果模型能够记住前两个非零数字,同时忽略掉一些不重要的干扰信息,那么即时序列很长,模型也有效地进行预测.
LSTM 模型在第 t 步时,循环单元的内部结构如图6.10所示.
提醒:为了和代码的实现保存一致性,这里使用形状为 (样本数量 × 序列长度 × 特征维度) 的张量来表示一组样本.
在本实验中,我们将使用第6.1.2.4节中定义Model_RNN4SeqClass模型,并构建 LSTM 算子.只需要实例化 LSTM 算,并传入Model_RNN4SeqClass模型,就可以用 LSTM 进行数字求和实验
LSTM层的代码与SRN层结构相似,只是在SRN层的基础上增加了内部状态、输入门、遗忘门和输出门的定义和计算。这里LSTM层的输出也依然为序列的最后一个位置的隐状态向量。代码实现如下:
import torch
import torch.nn as nn
import torch.nn.functional as F
class LSTM(nn.Module):
def __init__(self, input_size, hidden_size, Wi_attr=None, Wf_attr=None, Wo_attr=None, Wc_attr=None,
Ui_attr=None, Uf_attr=None, Uo_attr=None, Uc_attr=None, bi_attr=None, bf_attr=None,
bo_attr=None, bc_attr=None):
super(LSTM, self).__init__()
self.input_size = input_size
self.hidden_size = hidden_size
W_i = torch.randn([input_size, hidden_size])
W_f = torch.randn([input_size, hidden_size])
W_o = torch.randn([input_size, hidden_size])
W_c = torch.randn([input_size, hidden_size])
U_i = torch.randn([hidden_size, hidden_size])
U_f = torch.randn([hidden_size, hidden_size])
U_o = torch.randn([hidden_size, hidden_size])
U_c = torch.randn([hidden_size, hidden_size])
b_i = torch.randn([1, hidden_size])
b_f = torch.randn([1, hidden_size])
b_o = torch.randn([1, hidden_size])
b_c = torch.randn([1, hidden_size])
self.W_i = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(W_i, dtype=torch.float32), gain=1.0))
# 初始化模型参数
self.W_f = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(W_f, dtype=torch.float32), gain=1.0))
self.W_o = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(W_o, dtype=torch.float32), gain=1.0))
self.W_c = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(W_c, dtype=torch.float32), gain=1.0))
self.U_i = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(U_i, dtype=torch.float32), gain=1.0))
self.U_f = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(U_f, dtype=torch.float32), gain=1.0))
self.U_o = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(U_o, dtype=torch.float32), gain=1.0))
self.U_c = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(U_c, dtype=torch.float32), gain=1.0))
self.b_i = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(b_i, dtype=torch.float32), gain=1.0))
self.b_f = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(b_f, dtype=torch.float32), gain=1.0))
self.b_o = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(b_o, dtype=torch.float32), gain=1.0))
self.b_c = torch.nn.Parameter(torch.nn.init.xavier_uniform_(torch.as_tensor(b_c, dtype=torch.float32), gain=1.0))
# 初始化状态向量和隐状态向量
def init_state(self, batch_size):
hidden_state = torch.zeros([batch_size, self.hidden_size])
cell_state = torch.zeros([batch_size, self.hidden_size])
return hidden_state, cell_state
# 定义前向计算
def forward(self, inputs, states=None):
# inputs: 输入数据,其shape为batch_size x seq_len x input_size
batch_size, seq_len, input_size = inputs.shape
# 初始化起始的单元状态和隐状态向量,其shape为batch_size x hidden_size
if states is None:
states = self.init_state(batch_size)
hidden_state, cell_state = states
# 执行LSTM计算,包括:输入门、遗忘门和输出门、候选内部状态、内部状态和隐状态向量
for step in range(seq_len):
# 获取当前时刻的输入数据step_input: 其shape为batch_size x input_size
step_input = inputs[:, step, :]
# 计算输入门, 遗忘门和输出门, 其shape为:batch_size x hidden_size
I_gate = F.sigmoid(torch.matmul(step_input, self.W_i) + torch.matmul(hidden_state, self.U_i) + self.b_i)
F_gate = F.sigmoid(torch.matmul(step_input, self.W_f) + torch.matmul(hidden_state, self.U_f) + self.b_f)
O_gate = F.sigmoid(torch.matmul(step_input, self.W_o) + torch.matmul(hidden_state, self.U_o) + self.b_o)
# 计算候选状态向量, 其shape为:batch_size x hidden_size
C_tilde = F.tanh(torch.matmul(step_input, self.W_c) + torch.matmul(hidden_state, self.U_c) + self.b_c)
# 计算单元状态向量, 其shape为:batch_size x hidden_size
cell_state = F_gate * cell_state + I_gate * C_tilde
# 计算隐状态向量,其shape为:batch_size x hidden_size
hidden_state = O_gate * F.tanh(cell_state)
return hidden_state
Wi_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.2], [0.1, 0.2]]))
Wf_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.2], [0.1, 0.2]]))
Wo_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.2], [0.1, 0.2]]))
Wc_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.2], [0.1, 0.2]]))
Ui_attr = torch.nn.Parameter(torch.tensor([[0.0, 0.1], [0.1, 0.0]]))
Uf_attr = torch.nn.Parameter(torch.tensor([[0.0, 0.1], [0.1, 0.0]]))
Uo_attr = torch.nn.Parameter(torch.tensor([[0.0, 0.1], [0.1, 0.0]]))
Uc_attr = torch.nn.Parameter(torch.tensor([[0.0, 0.1], [0.1, 0.0]]))
bi_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.1]]))
bf_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.1]]))
bo_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.1]]))
bc_attr = torch.nn.Parameter(torch.tensor([[0.1, 0.1]]))
lstm = LSTM(2, 2, Wi_attr=Wi_attr, Wf_attr=Wf_attr, Wo_attr=Wo_attr, Wc_attr=Wc_attr,
Ui_attr=Ui_attr, Uf_attr=Uf_attr, Uo_attr=Uo_attr, Uc_attr=Uc_attr,
bi_attr=bi_attr, bf_attr=bf_attr, bo_attr=bo_attr, bc_attr=bc_attr)
inputs = torch.tensor([[[1, 0]]], dtype=torch.float32)
hidden_state = lstm(inputs)
print(hidden_state)
这里我们可以将自己实现的SRN和Paddle框架内置的SRN返回的结果进行打印展示,实现代码如下。
batch_size, seq_len, input_size = 8, 20, 32
inputs = torch.randn([batch_size, seq_len, input_size])
# 设置模型的hidden_size
hidden_size = 32
torch_lstm = nn.LSTM(input_size, hidden_size)
self_lstm = LSTM(input_size, hidden_size)
self_hidden_state = self_lstm(inputs)
torch_outputs, (torch_hidden_state, torch_cell_state) = torch_lstm(inputs)
print("self_lstm hidden_state: ", self_hidden_state.shape)
print("torch_lstm outpus:", torch_outputs.shape)
print("torch_lstm hidden_state:", torch_hidden_state.shape)
print("torch_lstm cell_state:", torch_cell_state.shape)
可以看到,自己实现的LSTM由于没有考虑多层因素,因此没有层次这个维度,因此其输出shape为[8, 32]。同时由于在以上代码使用Paddle内置API实例化LSTM时,默认定义的是1层的单向SRN,因此其shape为[1, 8, 32],同时隐状态向量为[8,20, 32].
接下来,我们可以将自己实现的LSTM与Paddle内置的LSTM在输出值的精度上进行对比,这里首先根据Paddle内置的LSTM实例化模型(为了进行对比,在实例化时只保留一个偏置,将偏置bihbih设置为0),然后提取该模型对应的参数,进行参数分割后,使用相应参数去初始化自己实现的LSTM,从而保证两者在参数初始化时是一致的。
在进行实验时,首先定义输入数据inputs
,然后将该数据分别传入Paddle内置的LSTM与自己实现的LSTM模型中,最后通过对比两者的隐状态输出向量。代码实现如下:
import torch
torch.manual_seed(0)
# 这里创建一个随机数组作为测试数据,数据shape为batch_size x seq_len x input_size
batch_size, seq_len, input_size, hidden_size = 2, 5, 10, 10
inputs = torch.randn([batch_size, seq_len, input_size])
# 设置模型的hidden_size
# bih_attr = torch.nn.Parameter(torch.tensor(torch.zeros([4*hidden_size, ])))
torch_lstm = nn.LSTM(input_size, hidden_size, bias=True)
# 获取torch_lstm中的参数,并设置相应的paramAttr,用于初始化lstm
print(torch_lstm.weight_ih_l0.T.shape)
chunked_W = torch.split(torch_lstm.weight_ih_l0.T, 4, dim=-1)
chunked_U = torch.split(torch_lstm.weight_hh_l0.T, 4, dim=-1)
chunked_b = torch.split(torch_lstm.bias_hh_l0.T, 4, dim=-1)
Wi_attr = torch.nn.Parameter(torch.tensor(chunked_W[0].clone().detach().requires_grad_(True)))
Wf_attr = torch.nn.Parameter(torch.tensor(chunked_W[1].clone().detach().requires_grad_(True)))
Wc_attr = torch.nn.Parameter(torch.tensor(chunked_W[2].clone().detach().requires_grad_(True)))
Wo_attr = torch.nn.Parameter(torch.tensor(chunked_W[3].clone().detach().requires_grad_(True)))
Ui_attr = torch.nn.Parameter(torch.tensor(chunked_U[0].clone().detach().requires_grad_(True)))
Uf_attr = torch.nn.Parameter(torch.tensor(chunked_U[1].clone().detach().requires_grad_(True)))
Uc_attr = torch.nn.Parameter(torch.tensor(chunked_U[2].clone().detach().requires_grad_(True)))
Uo_attr = torch.nn.Parameter(torch.tensor(chunked_U[3].clone().detach().requires_grad_(True)))
bi_attr = torch.nn.Parameter(torch.tensor(chunked_b[0].clone().detach().requires_grad_(True)))
bf_attr = torch.nn.Parameter(torch.tensor(chunked_b[1].clone().detach().requires_grad_(True)))
bc_attr = torch.nn.Parameter(torch.tensor(chunked_b[2].clone().detach().requires_grad_(True)))
bo_attr = torch.nn.Parameter(torch.tensor(chunked_b[3].clone().detach().requires_grad_(True)))
self_lstm = LSTM(input_size, hidden_size, Wi_attr=Wi_attr, Wf_attr=Wf_attr, Wo_attr=Wo_attr, Wc_attr=Wc_attr,
Ui_attr=Ui_attr, Uf_attr=Uf_attr, Uo_attr=Uo_attr, Uc_attr=Uc_attr,
bi_attr=bi_attr, bf_attr=bf_attr, bo_attr=bo_attr, bc_attr=bc_attr)
# 进行前向计算,获取隐状态向量,并打印展示
self_hidden_state = self_lstm(inputs)
torch_outputs, (torch_hidden_state, _) = torch_lstm(inputs)
print("torch SRN:\n", torch_hidden_state.detach().numpy().squeeze(0))
print("self SRN:\n", self_hidden_state.detach().numpy())
可以看到,两者的输出基本是一致的。另外,还可以进行对比两者在运算速度方面的差异。代码实现如下:
import time
# 这里创建一个随机数组作为测试数据,数据shape为batch_size x seq_len x input_size
batch_size, seq_len, input_size = 8, 20, 32
inputs = torch.randn([batch_size, seq_len, input_size])
# 设置模型的hidden_size
hidden_size = 32
self_lstm = LSTM(input_size, hidden_size)
torch_lstm = nn.LSTM(input_size, hidden_size)
# 计算自己实现的SRN运算速度
model_time = 0
for i in range(100):
strat_time = time.time()
hidden_state = self_lstm(inputs)
# 预热10次运算,不计入最终速度统计
if i < 10:
continue
end_time = time.time()
model_time += (end_time - strat_time)
avg_model_time = model_time / 90
print('self_lstm speed:', avg_model_time, 's')
# 计算torch内置的SRN运算速度
model_time = 0
for i in range(100):
strat_time = time.time()
outputs, (hidden_state, cell_state) = torch_lstm(inputs)
# 预热10次运算,不计入最终速度统计
if i < 10:
continue
end_time = time.time()
model_time += (end_time - strat_time)
avg_model_time = model_time / 90
print('torch_lstm speed:', avg_model_time, 's')
可以看到,由于Paddle框架的LSTM底层采用了C++实现并进行优化,Paddle框架内置的LSTM运行效率远远高于自己实现的LSTM。
在本节实验中,我们将使用6.1.2.4的Model_RNN4SeqClass作为预测模型,不同在于在实例化时将传入实例化的LSTM层。
动手联系6.2 在我们手动实现的LSTM算子中,是逐步计算每个时刻的隐状态。请思考如何实现更加高效的LSTM算子
# 基于RNN实现数字预测的模型
class Model_RNN4SeqClass(nn.Module):
def __init__(self, model, num_digits, input_size, hidden_size, num_classes):
super(Model_RNN4SeqClass, self).__init__()
# 传入实例化的RNN层,例如SRN
self.rnn_model = model
# 词典大小
self.num_digits = num_digits
# 嵌入向量的维度
self.input_size = input_size
# 定义Embedding层
self.embedding = Embedding(num_digits, input_size)
# 定义线性层
self.linear = nn.Linear(hidden_size, num_classes)
def forward(self, inputs):
# 将数字序列映射为相应向量
inputs_emb = self.embedding(inputs)
# 调用RNN模型
hidden_state = self.rnn_model(inputs_emb)
# 使用最后一个时刻的状态进行数字预测
logits = self.linear(hidden_state)
return logits
本节将基于RunnerV3类进行训练,首先定义模型训练的超参数,并保证和简单循环网络的超参数一致. 然后定义一个train
函数,其可以通过指定长度的数据集,并进行训练. 在train
函数中,首先加载长度为length
的数据,然后实例化各项组件并创建对应的Runner,然后训练该Runner。同时在本节将使用4.5.4节定义的准确度(Accuracy)作为评估指标,代码实现如下:
import os
import random
import torch
import numpy as np
# 训练轮次
num_epochs = 500
# 学习率
lr = 0.001
# 输入数字的类别数
num_digits = 10
# 将数字映射为向量的维度
input_size = 32
# 隐状态向量的维度
hidden_size = 32
# 预测数字的类别数
num_classes = 19
# 批大小
batch_size = 8
# 模型保存目录
save_dir = "./"
# 可以设置不同的length进行不同长度数据的预测实验
def train(length):
print(f"\n====> Training LSTM with data of length {length}.")
np.random.seed(0)
random.seed(0)
torch.manual_seed(0)
# 加载长度为length的数据
data_path = f"D:/datasets/{length}"
train_examples, dev_examples, test_examples = load_data(data_path)
train_set, dev_set, test_set = DigitSumDataset(train_examples), DigitSumDataset(dev_examples), DigitSumDataset(test_examples)
train_loader = torch.utils.data.DataLoader(train_set, batch_size=batch_size)
dev_loader = torch.utils.data.DataLoader(dev_set, batch_size=batch_size)
test_loader = torch.utils.data.DataLoader(test_set, batch_size=batch_size)
# 实例化模型
base_model = LSTM(input_size, hidden_size)
model = Model_RNN4SeqClass(base_model, num_digits, input_size, hidden_size, num_classes)
# 指定优化器
optimizer = torch.optim.Adam(lr=lr, params=model.parameters())
# 定义评价指标
metric = Accuracy()
# 定义损失函数
loss_fn = torch.nn.CrossEntropyLoss()
# 基于以上组件,实例化Runner
runner = RunnerV3(model, optimizer, loss_fn, metric)
# 进行模型训练
model_save_path = os.path.join(save_dir, f"best_lstm_model_{length}.pdparams")
runner.train(train_loader, dev_loader, num_epochs=num_epochs, eval_steps=100, log_steps=100, save_path=model_save_path)
return runner
接下来,分别进行数据长度为10, 15, 20, 25, 30, 35的数字预测模型训练实验,训练后的runner
保存至runners
字典中。
# LSTM训练
lstm_runners = {}
lengths = [10, 15, 20, 25, 30, 35]
for length in lengths:
runner = train(length)
lstm_runners[length] = runner
[Train] epoch: 473/500, step: 18000/19000, loss: 0.35202
[Evaluate] dev score: 0.87000, dev loss: 0.46193
[Train] epoch: 476/500, step: 18100/19000, loss: 0.09771
[Evaluate] dev score: 0.91000, dev loss: 0.38287
[Evaluate] best accuracy performence has been updated: 0.90000 --> 0.91000
[Train] epoch: 478/500, step: 18200/19000, loss: 0.02467
[Evaluate] dev score: 0.89000, dev loss: 0.42026
[Train] epoch: 481/500, step: 18300/19000, loss: 0.01818
[Evaluate] dev score: 0.89000, dev loss: 0.42676
[Train] epoch: 484/500, step: 18400/19000, loss: 0.04383
[Evaluate] dev score: 0.89000, dev loss: 0.42994
[Train] epoch: 486/500, step: 18500/19000, loss: 0.02579
[Evaluate] dev score: 0.89000, dev loss: 0.43919
[Train] epoch: 489/500, step: 18600/19000, loss: 0.02788
[Evaluate] dev score: 0.88000, dev loss: 0.44569
[Train] epoch: 492/500, step: 18700/19000, loss: 0.05951
[Evaluate] dev score: 0.88000, dev loss: 0.43005
[Train] epoch: 494/500, step: 18800/19000, loss: 0.02288
[Evaluate] dev score: 0.88000, dev loss: 0.44650
[Train] epoch: 497/500, step: 18900/19000, loss: 0.02292
[Evaluate] dev score: 0.88000, dev loss: 0.45742
[Evaluate] dev score: 0.88000, dev loss: 0.44346
[Train] Training done!
分别画出基于LSTM的各个长度的数字预测模型训练过程中,在训练集和验证集上的损失曲线,代码实现如下:
# # 画出训练过程中的损失图
for length in lengths:
runner = lstm_runners[length]
fig_name = f"D:/datasets/images/6.11_{length}.pdf"
plot_training_loss(runner, fig_name, sample_step=100)
plot_training_loss():
import matplotlib.pyplot as plt
def plot_training_loss(runner, fig_name, sample_step):
plt.figure()
train_items = runner.train_step_losses[::sample_step]
train_steps = [x[0] for x in train_items]
train_losses = [x[1] for x in train_items]
plt.plot(train_steps, train_losses, color='#e4007f', label="Train loss")
dev_steps = [x[0] for x in runner.dev_losses]
dev_losses = [x[1] for x in runner.dev_losses]
plt.plot(dev_steps, dev_losses, color='#f19ec2', linestyle='--', label="Dev loss")
# 绘制坐标轴和图例
plt.ylabel("loss", fontsize='large')
plt.xlabel("step", fontsize='large')
plt.legend(loc='upper right', fontsize='x-large')
plt.savefig(fig_name)
plt.show()
L=10、15、20、25、30、35
图6.11展示了LSTM模型在不同长度数据集上进行训练后的损失变化,同SRN模型一样,随着序列长度的增加,训练集上的损失逐渐不稳定,验证集上的损失整体趋向于变大,这说明当序列长度增加时,保持长期依赖的能力同样在逐渐变弱. 同图6.5相比,LSTM模型在序列长度增加时,收敛情况比SRN模型更好。
SRN:
LSTM模型在不同长度数据集上进行训练后的损失变化,同SRN模型一样,随着序列长度的增加,训练集上的损失逐渐不稳定,验证集上的损失整体趋向于变大,当序列长度增加时,保持长期依赖的能力同样在逐渐变弱。在序列长度增加时,LSTM收敛情况比SRN模型更好,确率也高,LSTM是通过门控机制来解决SRN的长程依赖问题,即随着训练时间的加长以及网络层数的增多,很容易出现梯度爆炸或梯度消失,导致无法处理较长序列数据,从而无法获取长距离数据的信息。
使用测试数据对在训练过程中保存的最好模型进行评价,观察模型在测试集上的准确率. 同时获取模型在训练过程中在验证集上最好的准确率,实现代码如下:
lstm_dev_scores = []
lstm_test_scores = []
for length in lengths:
print(f"Evaluate LSTM with data length {length}.")
runner = lstm_runners[length]
# 加载训练过程中效果最好的模型
model_path = os.path.join(save_dir, f"best_lstm_model_{length}.pdparams")
runner.load_model(model_path)
# 加载长度为length的数据
data_path = f"./datasets/{length}"
train_examples, dev_examples, test_examples = load_data(data_path)
test_set = DigitSumDataset(test_examples)
test_loader = DataLoader(test_set, batch_size=batch_size)
# 使用测试集评价模型,获取测试集上的预测准确率
score, _ = runner.evaluate(test_loader)
lstm_test_scores.append(score)
lstm_dev_scores.append(max(runner.dev_scores))
for length, dev_score, test_score in zip(lengths, lstm_dev_scores, lstm_test_scores):
print(f"[LSTM] length:{length}, dev_score: {dev_score}, test_score: {test_score: .5f}")
Evaluate LSTM with data length 20.
Evaluate LSTM with data length 25.
Evaluate LSTM with data length 30.
Evaluate LSTM with data length 35.
[LSTM] length:10, dev_score: 0.77, test_score: 0.77000
[LSTM] length:15, dev_score: 0.8, test_score: 0.82000
[LSTM] length:20, dev_score: 0.73, test_score: 0.76000
[LSTM] length:25, dev_score: 0.4, test_score: 0.31000
[LSTM] length:30, dev_score: 0.88, test_score: 0.88000
[LSTM] length:35, dev_score: 0.89, test_score: 0.82000
接下来,将SRN和LSTM在不同长度的验证集和测试集数据上的准确率绘制成图片,以方面观察。
#绘制全部图
import matplotlib.pyplot as plt
plt.plot(lengths, lstm_dev_scores, '-o', color='#e8609b', label="LSTM Dev Accuracy")
plt.plot(lengths, lstm_test_scores,'-o', color='#000000', label="LSTM Test Accuracy")
#绘制坐标轴和图例
plt.ylabel("accuracy", fontsize='large')
plt.xlabel("sequence length", fontsize='large')
plt.legend(loc='lower left', fontsize='x-large')
fig_name = "D:/datasets/images/6.12.pdf"
plt.savefig(fig_name)
plt.show()
下图展示了LSTM模型与SRN模型在不同长度数据集上的准确度对比。随着数据集长度的增加,LSTM模型在验证集和测试集上的准确率整体也趋向于降低;同时LSTM模型的准确率显著高于SRN模型,表明LSTM模型保持长期依赖的能力要优于SRN模型.
随着数据集长度的增加,LSTM模型和SRN模型的准确率降低,但是LSTM模型的准确率显著高于SRN模型,说明LSTM模型保持长期依赖的能力要优于SRN模型。随着数据集长度的增加,LSTM模型和SRN模型在验证集和测试集上的准确率整体均趋向于降低;LSTM模型的准确率显著高于SRN模型。SRN随着数据集长度的增加,其准确率不断下降
LSTM模型通过门控机制控制信息的单元状态的更新,这里可以观察当LSTM在处理一条数字序列的时候,相应门和单元状态是如何变化的。首先需要对以上LSTM模型实现代码中,定义相应列表进行存储这些门和单元状态在每个时刻的向量。
# 声明LSTM和相关参数
class LSTM(nn.Module):
def __init__(self, input_size, hidden_size, Wi_attr=None, Wf_attr=None, Wo_attr=None, Wc_attr=None,
Ui_attr=None, Uf_attr=None, Uo_attr=None, Uc_attr=None, bi_attr=None, bf_attr=None,
bo_attr=None, bc_attr=None):
super(LSTM, self).__init__()
self.input_size = input_size
self.hidden_size = hidden_size
# 初始化模型参数
if Wi_attr==None:
Wi=torch.zeros(size=[input_size, hidden_size], dtype=torch.float32)
else:
Wi = torch.tensor(Wi_attr, dtype=torch.float32)
self.W_i = torch.nn.Parameter(Wi)
if Wf_attr==None:
Wf=torch.zeros(size=[input_size, hidden_size], dtype=torch.float32)
else:
Wf = torch.tensor(Wf_attr, dtype=torch.float32)
self.W_f = torch.nn.Parameter(Wf)
if Wo_attr==None:
Wo=torch.zeros(size=[input_size, hidden_size], dtype=torch.float32)
else:
Wo = torch.tensor(Wo_attr, dtype=torch.float32)
self.W_o =torch.nn.Parameter(Wo)
if Wc_attr==None:
Wc=torch.zeros(size=[input_size, hidden_size], dtype=torch.float32)
else:
Wc = torch.tensor(Wc_attr, dtype=torch.float32)
self.W_c = torch.nn.Parameter(Wc)
if Ui_attr==None:
Ui = torch.zeros(size=[hidden_size, hidden_size], dtype=torch.float32)
else:
Ui = torch.tensor(Ui_attr, dtype=torch.float32)
self.U_i = torch.nn.Parameter(Ui)
if Uf_attr == None:
Uf = torch.zeros(size=[hidden_size, hidden_size], dtype=torch.float32)
else:
Uf = torch.tensor(Uf_attr, dtype=torch.float32)
self.U_f = torch.nn.Parameter(Uf)
if Uo_attr == None:
Uo = torch.zeros(size=[hidden_size, hidden_size], dtype=torch.float32)
else:
Uo = torch.tensor(Uo_attr, dtype=torch.float32)
self.U_o = torch.nn.Parameter(Uo)
if Uc_attr == None:
Uc = torch.zeros(size=[hidden_size, hidden_size], dtype=torch.float32)
else:
Uc = torch.tensor(Uc_attr, dtype=torch.float32)
self.U_c = torch.nn.Parameter(Uc)
if bi_attr == None:
bi = torch.zeros(size=[1,hidden_size], dtype=torch.float32)
else:
bi = torch.tensor(bi_attr, dtype=torch.float32)
self.b_i = torch.nn.Parameter(bi)
if bf_attr == None:
bf = torch.zeros(size=[1,hidden_size], dtype=torch.float32)
else:
bf = torch.tensor(bf_attr, dtype=torch.float32)
self.b_f = torch.nn.Parameter(bf)
if bo_attr == None:
bo = torch.zeros(size=[1,hidden_size], dtype=torch.float32)
else:
bo = torch.tensor(bo_attr, dtype=torch.float32)
self.b_o = torch.nn.Parameter(bo)
if bc_attr == None:
bc = torch.zeros(size=[1,hidden_size], dtype=torch.float32)
else:
bc = torch.tensor(bc_attr, dtype=torch.float32)
self.b_c = torch.nn.Parameter(bc)
# 初始化状态向量和隐状态向量
def init_state(self, batch_size):
hidden_state = torch.zeros(size=[batch_size, self.hidden_size], dtype=torch.float32)
cell_state = torch.zeros(size=[batch_size, self.hidden_size], dtype=torch.float32)
return hidden_state, cell_state
# 定义前向计算
def forward(self, inputs, states=None):
# inputs: 输入数据,其shape为batch_size x seq_len x input_size
batch_size, seq_len, input_size = inputs.shape
# 初始化起始的单元状态和隐状态向量,其shape为batch_size x hidden_size
if states is None:
states = self.init_state(batch_size)
hidden_state, cell_state = states
# 定义相应的门状态和单元状态向量列表
self.Is = []
self.Fs = []
self.Os = []
self.Cs = []
# 初始化状态向量和隐状态向量
cell_state = torch.zeros(size=[batch_size, self.hidden_size], dtype=torch.float32)
hidden_state = torch.zeros(size=[batch_size, self.hidden_size], dtype=torch.float32)
# 执行LSTM计算,包括:隐藏门、输入门、遗忘门、候选状态向量、状态向量和隐状态向量
for step in range(seq_len):
input_step = inputs[:, step, :]
I_gate = F.sigmoid(torch.matmul(input_step, self.W_i) + torch.matmul(hidden_state, self.U_i) + self.b_i)
F_gate = F.sigmoid(torch.matmul(input_step, self.W_f) + torch.matmul(hidden_state, self.U_f) + self.b_f)
O_gate = F.sigmoid(torch.matmul(input_step, self.W_o) + torch.matmul(hidden_state, self.U_o) + self.b_o)
C_tilde = F.tanh(torch.matmul(input_step, self.W_c) + torch.matmul(hidden_state, self.U_c) + self.b_c)
cell_state = F_gate * cell_state + I_gate * C_tilde
hidden_state = O_gate * F.tanh(cell_state)
# 存储门状态向量和单元状态向量
self.Is.append(I_gate.detach().numpy().copy())
self.Fs.append(F_gate.detach().numpy().copy())
self.Os.append(O_gate.detach().numpy().copy())
self.Cs.append(cell_state.detach().numpy().copy())
return hidden_state
接下来,需要使用新的LSTM模型,重新实例化一个runner,本节使用序列长度为10的模型进行此项实验,因此需要加载序列长度为10的模型。
# 实例化模型
base_model = LSTM(input_size, hidden_size)
model = Model_RNN4SeqClass(base_model, num_digits, input_size, hidden_size, num_classes)
# 指定优化器
optimizer = torch.optim.Adam(lr=lr, params=model.parameters())
# 定义评价指标
metric = Accuracy()
# 定义损失函数
loss_fn = torch.nn.CrossEntropyLoss()
# 基于以上组件,重新实例化Runner
runner = RunnerV3(model, optimizer, loss_fn, metric)
length = 10
# 加载训练过程中效果最好的模型
model_path = os.path.join(save_dir, f"best_lstm_model_{length}.pdparams")
runner.load_model(model_path)
接下来,给定一条数字序列,并使用数字预测模型进行数字预测,这样便会将相应的门状态和单元状态向量保存至模型中. 然后分别从模型中取出这些向量,并将这些向量进行绘制展示。代码实现如下:
import seaborn as sns
import matplotlib.pyplot as plt
def plot_tensor(inputs, tensor, save_path, vmin=0, vmax=1):
tensor = np.stack(tensor, axis=0)
tensor = np.squeeze(tensor, 1).T
plt.figure(figsize=(16,6))
# vmin, vmax定义了色彩图的上下界
ax = sns.heatmap(tensor, vmin=vmin, vmax=vmax)
ax.set_xticklabels(inputs)
ax.figure.savefig(save_path)
# 定义模型输入
inputs = [6, 7, 0, 0, 1, 0, 0, 0, 0, 0]
X = torch.as_tensor(inputs.copy())
X = X.unsqueeze(0)
# 进行模型预测,并获取相应的预测结果
logits = runner.predict(X)
predict_label = torch.argmax(logits, dim=-1)
print(f"predict result: {predict_label.numpy()[0]}")
# 输入门
Is = runner.model.rnn_model.Is
plot_tensor(inputs, Is, save_path="./images/6.13_I.pdf")
# 遗忘门
Fs = runner.model.rnn_model.Fs
plot_tensor(inputs, Fs, save_path="./images/6.13_F.pdf")
# 输出门
Os = runner.model.rnn_model.Os
plot_tensor(inputs, Os, save_path="./images/6.13_O.pdf")
# 单元状态
Cs = runner.model.rnn_model.Cs
plot_tensor(inputs, Cs, save_path="./images/6.13_C.pdf", vmin=-5, vmax=5)
图6.13 当LSTM处理序列数据[6, 7, 0, 0, 1, 0, 0, 0, 0, 0]的过程中单元状态和门数值的变化图,其中横坐标为输入数字,纵坐标为相应门或单元状态向量的维度,颜色的深浅代表数值的大小。可以看到,当输入门遇到不同位置的数字0时,保持了相对一致的数值大小,表明对于0元素保持相同的门控过滤机制,避免输入信息的变化给当前模型带来困扰;当遗忘门遇到数字1后,遗忘门数值在一些维度上变小,表明对某些信息进行了遗忘;随着序列的输入,输出门和单元状态在某些维度上数值变小,在某些维度上数值变大,表明输出门在根据信息的重要性选择信息进行输出,同时单元状态也在保持着对文本预测重要的一些信息.
色阶图中,横坐标为输入数字,纵坐标为相应门或单元状态向量的维度,颜色的深浅表示数值的大小。可以看到,当输入门遇到0时,保持了数值大小基本不变,表明对于0元素保持相同的门控过滤机制,避免输入信息的变化给当前模型带来困扰;当遗忘门遇到数字1后,遗忘门数值在一些维度上变小,表明对某些信息进行了遗忘;随着序列的输入,输出门和单元状态在某些维度上数值变小,在某些维度上数值变大,表明输出门在根据信息的重要性选择信息进行输出,同时单元状态也在保持着对文本预测重要的一些信息.
个人总结 :通过本次实验LSTM的原理和LSTM记忆能力,对SRN和LSTM也进行了对比,发现了LSTM相对于SRN对于时间序列的强大的记忆功能对于LSTM的理解更加深刻了有很大收获。
ref:
https://blog.csdn.net/qq_38975453/article/details/126800091
NNDL 实验6(上) - HBU_DAVID - 博客园 (cnblogs.com)
8. 循环神经网络 — 动手学深度学习 2.0.0-beta1 documentation (d2l.ai)
9. 现代循环神经网络 — 动手学深度学习 2.0.0-beta1 documentation (d2l.ai)