- 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
- 🍖 原作者:K同学啊
一、前期准备工作
import torch.nn.functional as F
import numpy as np
import pandas as pd
import torch
from torch import nn
1. 导入数据
data = pd.read_csv(r"/home/aiusers/space_yjl/深度学习训练营/进阶/第R4周:LSTM-火灾温度预测/woodpine2.csv")
data
2. 数据集可视化
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np # 新增导入numpy库
# 读取CSV文件
data = pd.read_csv(r'/home/aiusers/space_yjl/深度学习训练营/进阶/第R4周:LSTM-火灾温度预测/woodpine2.csv')
# 提取列数据,并转换为numpy数组
time = np.array(data['Time'])
tem1 = np.array(data['Tem1'])
co1 = np.array(data['CO 1'])
soot1 = np.array(data['Soot 1'])
# 绘制折线图
plt.plot(time, tem1, label='Tem1')
plt.plot(time, co1, label='CO 1')
plt.plot(time, soot1, label='Soot 1')
# 添加标题和坐标轴标签
plt.title('Data Visualization')
plt.xlabel('Time')
plt.ylabel('Values')
# 添加图例
plt.legend()
# 显示图形
plt.show()
训练营中的这个有点问题 未解决 我换了一种方式可视化数据
import matplotlib.pyplot as plt
import seaborn as sns
plt.rcParams['savefig.dpi'] = 500 #图片像素
plt.rcParams['figure.dpi'] = 500 #分辨率
fig, ax =plt.subplots(1,3,constrained_layout=True, figsize=(14, 3))
sns.lineplot(data=data["Tem1"], ax=ax[0])
sns.lineplot(data=data["CO 1"], ax=ax[1])
sns.lineplot(data=data["Soot 1"], ax=ax[2])
plt.show()
dataFrame = data.iloc[:,1:]
dataFrame
二、构建数据集
1. 数据集预处理
from sklearn.preprocessing import MinMaxScaler
dataFrame = data.iloc[:,1:].copy()
sc = MinMaxScaler(feature_range=(0, 1)) #将数据归一化,范围是0到1
for i in ['CO 1', 'Soot 1', 'Tem1']:
dataFrame[i] = sc.fit_transform(dataFrame[i].values.reshape(-1, 1))
dataFrame.shape
2. 设置X、y
width_X = 8
width_y = 1
##取前8个时间段的Tem1、CO 1、Soot 1为X,第9个时间段的Tem1为y。
X = []
y = []
in_start = 0
for _, _ in data.iterrows():
in_end = in_start + width_X
out_end = in_end + width_y
if out_end < len(dataFrame):
X_ = np.array(dataFrame.iloc[in_start:in_end , ])
y_ = np.array(dataFrame.iloc[in_end :out_end, 0])
X.append(X_)
y.append(y_)
in_start += 1
X = np.array(X)
y = np.array(y).reshape(-1,1,1)
X.shape, y.shape
检查数据集中是否有空值
print(np.any(np.isnan(X)))
print(np.any(np.isnan(y)))
3. 划分数据集
X_train = torch.tensor(np.array(X[:5000]), dtype=torch.float32)
y_train = torch.tensor(np.array(y[:5000]), dtype=torch.float32)
X_test = torch.tensor(np.array(X[5000:]), dtype=torch.float32)
y_test = torch.tensor(np.array(y[5000:]), dtype=torch.float32)
X_train.shape, y_train.shape
from torch.utils.data import TensorDataset, DataLoader
train_dl = DataLoader(TensorDataset(X_train, y_train),
batch_size=64,
shuffle=False)
test_dl = DataLoader(TensorDataset(X_test, y_test),
batch_size=64,
shuffle=False)
三、模型训练
1. 构建模型
class model_lstm(nn.Module):
def __init__(self):
super(model_lstm, self).__init__()
self.lstm0 = nn.LSTM(input_size=3 ,hidden_size=320,
num_layers=1, batch_first=True)
self.lstm1 = nn.LSTM(input_size=320 ,hidden_size=320,
num_layers=1, batch_first=True)
self.fc0 = nn.Linear(320, 1)
def forward(self, x):
out, hidden1 = self.lstm0(x)
out, _ = self.lstm1(out, hidden1)
out = self.fc0(out)
return out[:, -1:, :] #取2个预测值,否则经过lstm会得到8*2个预测
model = model_lstm()
model
model_lstm(
(lstm0): LSTM(3, 320, batch_first=True)
(lstm1): LSTM(320, 320, batch_first=True)
(fc0): Linear(in_features=320, out_features=1, bias=True)
)
模型的输出数据集格式是什么
model(torch.rand(30,8,3)).shape
2. 定义训练函数
# 训练循环
import copy
def train(train_dl, model, loss_fn, opt, lr_scheduler=None):
size = len(train_dl.dataset)
num_batches = len(train_dl)
train_loss = 0 # 初始化训练损失和正确率
for x, y in train_dl:
x, y = x.to(device), y.to(device)
# 计算预测误差
pred = model(x) # 网络输出
loss = loss_fn(pred, y) # 计算网络输出和真实值之间的差距
# 反向传播
opt.zero_grad() # grad属性归零
loss.backward() # 反向传播
opt.step() # 每一步自动更新
# 记录loss
train_loss += loss.item()
if lr_scheduler is not None:
lr_scheduler.step()
print("learning rate = {:.5f}".format(opt.param_groups[0]['lr']), end=" ")
train_loss /= num_batches
return train_loss
3. 定义测试函数
def test (dataloader, model, loss_fn):
size = len(dataloader.dataset) # 测试集的大小
num_batches = len(dataloader) # 批次数目
test_loss = 0
# 当不进行训练时,停止梯度更新,节省计算内存消耗
with torch.no_grad():
for x, y in dataloader:
x, y = x.to(device), y.to(device)
# 计算loss
y_pred = model(x)
loss = loss_fn(y_pred, y)
test_loss += loss.item()
test_loss /= num_batches
return test_loss
4. 正式训练模型
#设置GPU训练
device=torch.device("cuda" if torch.cuda.is_available() else "cpu")
device
#训练模型
model = model_lstm()
model = model.to(device)
loss_fn = nn.MSELoss() # 创建损失函数
learn_rate = 1e-1 # 学习率
opt = torch.optim.SGD(model.parameters(),lr=learn_rate,weight_decay=1e-4)
epochs = 50
train_loss = []
test_loss = []
lr_scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(opt,epochs, last_epoch=-1)
for epoch in range(epochs):
model.train()
epoch_train_loss = train(train_dl, model, loss_fn, opt, lr_scheduler)
model.eval()
epoch_test_loss = test(test_dl, model, loss_fn)
train_loss.append(epoch_train_loss)
test_loss.append(epoch_test_loss)
template = ('Epoch:{:2d}, Train_loss:{:.5f}, Test_loss:{:.5f}')
print(template.format(epoch+1, epoch_train_loss, epoch_test_loss))
print("="*20, 'Done', "="*20)
四、模型评估
1. LOSS图
import matplotlib.pyplot as plt
plt.figure(figsize=(5, 3),dpi=120)
plt.plot(train_loss , label='LSTM Training Loss')
plt.plot(test_loss, label='LSTM Validation Loss')
plt.title('Training and Validation Loss')
plt.legend()
plt.show()
2. 调用模型进行预测
predicted_y_lstm = sc.inverse_transform(model(X_test).detach().numpy().reshape(-1,1)) # 测试集输入模型进行预测
y_test_1 = sc.inverse_transform(y_test.reshape(-1,1))
y_test_one = [i[0] for i in y_test_1]
predicted_y_lstm_one = [i[0] for i in predicted_y_lstm]
plt.figure(figsize=(5, 3),dpi=120)
# 画出真实数据和预测数据的对比曲线
plt.plot(y_test_one[:2000], color='red', label='real_temp')
plt.plot(predicted_y_lstm_one[:2000], color='blue', label='prediction')
plt.title('Title')
plt.xlabel('X')
plt.ylabel('Y')
plt.legend()
plt.show()
3. R2值评估
from sklearn import metrics
"""
RMSE :均方根误差 -----> 对均方误差开方
R2 :决定系数,可以简单理解为反映模型拟合优度的重要的统计量
"""
RMSE_lstm = metrics.mean_squared_error(predicted_y_lstm_one, y_test_1)**0.5
R2_lstm = metrics.r2_score(predicted_y_lstm_one, y_test_1)
print('均方根误差: %.5f' % RMSE_lstm)
print('R2: %.5f' % R2_lstm)
五、个人总结
总的来说,这个模型先通过两层 LSTM 层对输入的序列数据进行特征提取和对序列中信息的记忆、传递处理,然后利用全连接层将提取到的特征转换为具体的预测值,并筛选出最后时刻对应的预测输出。
输入形状
对于这个模型,输入数据x的形状假设为(batch_size, seq_length, input_size)。在代码中,input_size被定义为3,batch_size是每次输入的批量大小(在代码中未明确限制,但由数据加载等环节决定),seq_length是序列长度,即每个样本序列包含的时间步数。
中间层形状变化
第一个 LSTM 层(self.lstm0):
输入形状为(batch_size, seq_length, 3),经过self.lstm0后,输出out的形状变为(batch_size, seq_length, 320),隐藏状态hidden1的形状为(1, batch_size, 320)。这里的320是第一个 LSTM 层定义的隐藏状态维度hidden_size,1是层数(num_layers)。
第二个 LSTM 层(self.lstm1):
它接收第一个 LSTM 层的输出out(形状为(batch_size, seq_length, 320))和隐藏状态hidden1(形状为(1, batch_size, 320))作为输入。输出out的形状在经过self.lstm1后仍然保持为(batch_size, seq_length, 320),因为这一层的参数设置(input_size = 320,hidden_size = 320)没有改变数据的维度规模,只是对数据进行了进一步的特征提取和序列信息处理。
全连接层(self.fc0):
接收形状为(batch_size, seq_length, 320)的out,经过线性变换self.fc0后,输出out的形状变为(batch_size, seq_length, 1)。这是因为全连接层的定义是nn.Linear(320, 1),将输入的维度为320的特征向量映射到维度为1的输出空间。
输出形状
最后通过return out[:, -1:, :]操作,输出形状变为(batch_size, 1, 1)。这里的操作是提取每个样本序列(batch_size个样本)的最后一个时间步(-1:表示最后一个位置)对应的预测值,并且由于全连接层输出的最后一个维度是1,所以最终输出形状是(batch_size, 1, 1),每个样本对应一个最终的预测值(维度为1)。